Multimodal intent detection leverages complementary information from diverse sensors to achieve precise semantic understanding; however, existing methodologies predominantly operate under the ideal assumption of modality completeness. In practical deployments, missing modalities, stemming from sensor failure, background noise, or privacy constraints, are inevitable and uncertain, leading to severe performance degradation. To address this challenge, we propose SADF, a Structure Aware Distillation Framework designed to facilitate robust cross-level knowledge transfer from a full-modality teacher to a partial-modality student. At the semantic level, semantic topology distillation aligns prototype similarity distributions between the teacher and the student, capturing class topology and hierarchical relations. At the discriminative level, boundary aware distillation decouples predictions into target and nontarget classes, enforcing alignment with the teacher to enhance discriminability and suppress noise. By integrating these strategies, SADF enables the student to learn robust representations and decisions when uncertain modalities are missing. The results of experiments based on two benchmarks demonstrate that SADF consistently outperforms strong baselines.

Mobile Crowdsensing (MCS) is a new data acquisition method that has emerged with the proliferation of smart mobile devices. With the expanding scale of urban sensing, the locations of tasks and users become critical information, which plays a significant role in crowd-sensing and task scheduling areas. Tasks in areas with a high concentration of users can be completed quickly, whereas tasks in sparsely populated areas are challenging to accomplish. To address this issue, existing research has primarily focused on task assignment to designated users, assuming that users' motivations are rational, while neglecting the impact of psychological factors on their motivations. Therefore, we propose an incentive mechanism based on prospect theory, analyzing the decisions users might make under irrationality and then adjusting corresponding rewards to influence user decisions. This paper transforms the problem of maximizing the data value in crowdsensing into an ordered submodular function model. Our proposed incentive mechanism consists of three components: User Decision-Making, User Selection, and Payment Determination. In the User Decision-Making phase, users calculate the prospect value based on the auction results from the previous round to make decisions. In the User Selection phase, users are chosen based on marginal value. In the Payment Determination phase, rewards for winning users are designed based on the ordered submodular model. The platform provides auction results as a reference for the next round. In the experimental section, we demonstrate that the incentive mechanism can enhance the platform's value.

With the ever-growing attention on communication security, machine learning-based network intrusion detection system (NIDS) is widely utilized to meet different security requirements. However, most of the existing methods manually extract or learn features from raw traffic, which is usually expensive, complicated, and time-consuming. Moreover, this also brings unprecedented challenges for preserving users’ privacy in the communication process, making it difficult for existing solutions to be deployed in practice due to the privacy requirements from legal policies. This paper proposes a privacy-preserving graph neural network (named NIGNN) for NIDS, which can encode the local structure and traffic features. To address the privacy issues pertaining to the application of graph representation learning, we design a privacy message-passing mechanism with formal privacy guarantees, in which sensitive information potentially contained in graph vertices will be kept private. Specifically, we design a privacy-enhancement graph representation that introduces a degree-sensitive item in vertex-based aggregation to reduce noise. Our theoretical analysis shows that NIGNN can provide a provable privacy guarantee. Extensive experiments demonstrate NIGNN's performance in maintaining a sound privacy-accuracy trade-off.

Model stealing (MS) involves querying and observing the output of a machine learning model to steal its capabilities. The quality of queried data is crucial, yet obtaining a large amount of real data for MS is often challenging. Recent works have reduced reliance on real data by using generative models. However, when high-dimensional query data is required, these methods are impractical due to the high costs of querying and the risk of model collapse. In this work, we propose using sample gradients (SG) to enhance the utility of each real sample, as SG provides crucial guidance on the decision boundaries of the victim model. However, utilizing SG in the model stealing scenario faces two challenges: 1. Pixel-level gradient estimation requires ex-tensive query volume and is susceptible to defenses. 2. The estimation of sample gradients has a significant variance. This paper proposes Superpixel Sample Gradient stealing (SPSG) for model stealing under the constraint of limited real samples. With the basic idea of imitating the victim model's low-variance patch-level gradients instead ofpixel-level gradients, SPSG achieves efficient sample gradient es-timation through two steps. First, we perform patch-wise perturbations on query images to estimate the average gradient in different regions of the image. Then, we filter the gradients through a threshold strategy to reduce variance. Exhaustive experiments demonstrate that, with the same number of real samples, SPSG achieves accuracy, agreements, and adversarial success rate significantly surpassing the current state-of-the-art MS methods. Codes are available at https://github.com/zyI123456aBISPSG_attack.

With the ever-growing interest in modeling complex graph structures, graph neural networks (GNN) provide a generalized form of exploiting non-Euclidean space data. However, the global graph may be distributed across multiple data centers, which makes conventional graph-based models incapable of modeling a complete graph structure. This also brings an unprecedented challenge to user privacy protection in distributed graph learning. Due to privacy requirements of legal policies, existing graph-based solutions are difficult to deploy in practice. In this paper, we propose a privacy-preserving graph neural network based on local graph augmentation, named LGA-PGNN, which preserves user privacy by enforcing local differential privacy (LDP) noise into the decentralized local graphs held by different data holders. Moreover, we perform local neighborhood augmentation on low-degree vertices to enhance the expressiveness of the learned model. Specifically, we propose two graph privacy attacks, namely attribute inference attack and link stealing attack, which aim at compromising user privacy. The experimental results demonstrate that LGA-PGNN can effectively mitigate these two attacks and provably avoid potential privacy leakage while ensuring the utility of the learning model.

The Internet of Things (IoT) mainly consists of a large number of Internet-connected devices. The proliferation of untrusted third-party IoT applications has led to an increase in IoT-based malware attacks. In addition, it is infeasible for the IoT devices to support the sophisticated detection systems due to the restricted resources. Edge computing is considered to be promising. It provides solutions to the data security and privacy leakage brought by untrusted third-party IoT applications. In this article, an intelligent trusted and secure edge computing (ITEC) system is proposed for IoT malware detection. In this system, a signature-based preidentification mechanism is built for matching and identifying the malicious behaviors of untrusted third-party IoT applications. A delay strategy is then embedded into the risk detection engine in order to “buy time” for threat analysis and rate-limit the impact of suspicious third-party IoT applications in the system. We conduct extensive experiments to verify the effectiveness of the ITEC system and show that we can achieve accuracies of up to 98.52%.

Graph Neural Networks (GNNs) as an emerging technique have shown excellent performance in a variety of fields, such as social networks and recommendation systems. However, GNNs may have to overcome privacy concerns as large amounts of information about their training datasets may be compromised. In this paper, we develop a privacy-preserving GNN to enforce privacy preservation, which utilizes a private Functional Mechanism (FM) to train the learning model. This mechanism perturbs the polynomial approximation of the objective function to enforce Differential Privacy (DP) in the GNN model. We show that our method can maximize the accuracy of the results with comparable prediction power to the unperturbed results while satisfying the privacy guarantees.

Neural Networks’ performance on a sequence of incremental class tasks drops over time for Class Incremental Learning (IL). The gradient-based IL methods can simultaneously adapt to both new and previous tasks by promoting the update of model in the correct direction. However, existing methods simply consider the previous/new task gradients separately. In this paper, we propose Parallel Gradient Blend (PGB) paradigm. On the one hand, PGB uses the gradients generated by mixing previous and new samples in equal proportions with Batch-Normal layers to adjust a reasonable model update direction. By comparing gradient similarities, the model selects either the previous task gradient or mixed gradient to update. On the other hand, PGB uses the sample feature gradient distribution difference to construct a regularized gradient. Finally, we experimentally demonstrate that PGB outperforms state-of-the-art methods on class-IL benchmarks.

As the demand for Internet of Things (IoT) technologies continues to grow, IoT devices have been viable targets for malware infections. Although deep learning-based malware detection has achieved great success, the detection models are usually trained based on the collected user records, thereby leading to significant privacy risks. One promising solution is to leverage federated learning (FL) to enable distributed on-device training without centralizing the private user records. However, it is non-trivial for IoT users to label these records, where the quality and the trustworthiness of data labeling are hard to guarantee. To address the above issues, this paper develops a semi-supervised federated IoT malware detection framework based on knowledge transfer technologies, named by FedMalDE. Specifically, FedMalDE explores the underlying correlation between labeled and unlabeled records to infer labels towards unlabeled samples by the knowledge transfer mechanism. Moreover, a specially designed subgraph aggregated capsule network (SACN) is used to efficiently capture varied malicious behaviors. The extensive experiments conducted on real-world data demonstrate the effectiveness of FedMalDE in detecting IoT malware and its sufficient privacy and robustness guarantee.

With the coming of the Internet of Things (IoT) era, malware attacks targeting IoT networks have posed serious threats to users. Recently, the emerging of edge computing have paved the way for new data processing paradigms in IoT networks, but it is still a challenge for deploying malware detection systems on the IoT devices. This paper develops an IoT malware detection system based on trust hybrid user-edge evaluation, namely MDHE. This system decomposes a large and complex deep learning model into two parts, which are deployed on edge servers and end devices, respectively. Specifically, a trust evaluation mechanism is used to select the trusted devices to participate the model training. Moreover, we develop a private feature generation that leverages a graph mining technology to extract the subgraph features, which then are perturbed by leveraging the differential privacy technology to prevent user privacy from leaking. Finally, we reconstruct the perturbed features on edge server, and propose a Capsule Network (CapsNet) to identify malware. Experimental results show that MDHE can effectively detect malware. Specifically, it can reduce sensitive inference while maintaining the utility of data.

The advent of 5G brought new opportunities to leapfrog beyond current Industrial Internet of Things (IoT). However, the ever-growing IoT has also attracted adversaries to develop new malware attacks against various IoT applications. Although deep-learning-based methods are expected to combat the sophisticated malwares by exploring the latent attack patterns, such detection can be hardly supported by battery-powered end devices, such as Android-based smartphones. Edge computing enables the near-real-time analysis of IoT data by migrating artificial intelligence (AI)-enabled computation-intensive tasks from resource-constrained IoT devices to nearby edge servers. However, owing to varying channel conditions and the demanding latency requirements of malware detection, it is challenging to coordinate the computing task offloading among multiple users. By leveraging the computation capacity and the proximity benefits of edge computing, we propose a hierarchical security framework for IoT malware detection. Considering the complexity of the AI-enabled malware detection task, we provide a delay-aware computational offloading strategy with minimum delay. Specifically, we construct a coordinated representation learning model, named by Two-Stream Attention-Caps, to capture the latent behavioral patterns of evolving malware attacks. Experimental results show that our system consistently outperforms the state-of-the-art systems in detection performance on four benchmark datasets.

Given the distributed nature of the massively connected “Things” in IoT, IoT networks have been a primary target for cyberattacks. Although machine learning based network intrusion detection systems (NIDS) can effectively detect abnormal network traffic behaviors, most existing approaches are based on a large amount of labeled traffic flow data, which hinders their implementation in the highly dynamic IoT networks with limited labeling. In this paper, we develop a novel Flow Topology based Graph Convolutional Network (FT-GCN) approach for label-limited IoT network intrusion detection. Our main idea is to leverage the underlying traffic flow patterns,$i.e.$, the flow topological structure, to unlock the full potential of the traffic flow data with limited labeling, where the FT-GCN will be deployed at the edge servers in IoT networks to detect intrusions via software defined network technologies. Specifically, FT-GCN first takes the time correlation of traffic flows into account to construct an interval-constrained traffic graph (ICTG). Besides, a Node-Level Spatial (NLS) attention mechanism is designed to further enhance the key statistical features of traffic flows in ICTG. Finally, the combined representation of statistical flow features and flow topological structure are learned by the cost-effective Topology Adaptive Graph Convolutional Networks (TAGCN) for intrusion identification in IoT networks. Extensive experiments are conducted on three real-world datasets, which demonstrate the effectiveness of the proposed FT-GCN compared to state-of-the-art approaches.