With the advent of cloud computing, traditional machine learning (ML) are migrating into cloud-based ML day by day following the concept of machine learning as a cloud service, which enables multiple entities to contribute to and benefit from shared datasets and models. As well as training the linear classification model, training the nonlinear classification model is also an essential task in cloud-based ML. However, this task commonly involves learning knowledge from different datasets provided by various entities, which often contain sensitive information like patients' physiological indices. Therefore, it gives rise a natural question how to allow multiple users collaboratively participating in a nonlinear classification task while preserving these datas' privacy. As a promising cryptographic tool, the concept of unbounded multiinput functional encryption can be developed to answer such a question, such as google search engines are running over this concept-based ML approaches. However, most of existing approaches are derived from this concept with inner product functionality, specifying for a linear classification model and thus fails to cope with a non-linear classification one. In this paper, we introduce an advanced cryptographic concept called unbounded multi-input quadratic functional encryption, and give a concrete construction which allows arbitrary number of users participating in the classifying tasks with a nonlinear classification model but without divulging their private data. Moreover, we provide a strict mathematical security proof under a well-defined security model as well as some security attacks are analyzed, followed by an experimental analysis and comparison on a real dateset as well as a practical use case to demonstrate our scheme's performance.

Human action recognition (HAR), which aims to recognize and understand individual actions and intentions, has rapidly become a research hotspot in computer vision. Compared with other data modalities, skeleton data offers more efficient node semantics and more coherent spatio-temporal motion patterns, effectively reducing the impact of lighting and background changes. In recent years, many researchers have focused on skeleton-based action recognition methods and have made significant progress. However, we believe that the current skeleton-based action recognition methods still face three major challenges: 1) reducing reliance on expensive labeled data while maintaining model performance; 2) enabling the model to understand and recognize new behavior classes with a limited number of samples; and 3) addressing the challenges posed by the lack of skeleton information in single-modality spatio-temporal motion representation learning. Based on these challenges, we conduct a comprehensive review of the existing skeleton-based action recognition methods. Additionally, we provide an extensive review and analysis of publicly available action recognition datasets. This review aims to offer researchers a comprehensive perspective, stimulate more innovative ideas, and promote the application and breakthrough of skeleton action recognition in a wider range of computer vision tasks.

As public awareness of privacy protection increases and data become more valuable, the applications of federated learning (FL) in the emerging field of Artificial Internet of Things (AIoT) has received widespread attention. Meanwhile, differential privacy (DP), providing strict privacy guarantees, has been introduced to meet users’ stringent privacy protection needs and increasingly sound laws and regulations. However, the implementations of DP in multiple iterations and rounds of FL training, as well as adding noise to all parameters without differentiation, will cause noise accumulation, resulting in the FL system to decline in performance or even fail to converge. To address the issue, FL with denoising DP and sparsification (DDPS-FL) is proposed in this article. First, a local denoising mechanism (LDM) suitable for DP with arbitrary noise adding mechanism is proposed. By removing the previously added noise from global models, LDM achieves direct noise reduction for clients. Second, sparsification based on parameter variation (SPV) is proposed to reduce noise indirectly by deleting nonsignificant parameters without compromising the level of privacy protection. Besides, SPV is able to multiple beneficial effects, such as saving privacy budget, stimulating the dynamism of FL training, amplifying privacy protection effect, and improving communication efficiency. Third, theoretical analysis is performed to prove that DDPS-FL can guarantee user privacy and has ideal convergence, and to analyze the impact of parameters, such as the number of training rounds and iterations on the system performance. Evaluation experiments based on four real-world datasets are elaborated to show that DDPS-FL outperforms state-of-the-art schemes in terms of training stability, model accuracy, and communication efficiency, and its performance becomes relatively better when more noise is added.

Many privacy-preserving multidimensional data aggregation (PPMDA) schemes have been proposed to safeguard user privacy and provide aggregated real-time data for the control center (CC) to optimize power allocation in the smart grid. However, without Trusted Authority (TA), existing PPMDA schemes cannot simultaneously provide lightweight encryption, achieve fault tolerance, and resist collusion attacks between fog nodes and CC. To address these issues, we propose a fault-tolerant and privacy-enhanced multidimensional data aggregation scheme without TA (FPMDA) based on fog computing. Specifically, the Chinese Remainder Theorem is leveraged to enhance the efficiency of multidimensional data processing, and an innovative dual-masking approach is introduced to ensure the security of data aggregation. In addition, employing the homomorphic property of the (t, k)-threshold secret sharing algorithm, we design a data aggregation method that enhances security and fault tolerance, making it resilient against insider attacks. Finally, compared with existing schemes, FPMDA not only significantly enhances privacy preservation while maintaining required security properties but also achieves low computational and communication load, demonstrating practicality for resource-constrained smart meters.

Autonomous vehicles (AVs) rely on controller area network (CAN), which ensures the communication between massive electronic control units (ECUs) and passenger safety. Although CAN is a lightweight and reliable broadcast protocol, its vulnerability has caused CAN to confront serious security threats. Adversaries and malicious organizations can impair CAN bus in a variety of ways, such as injecting malicious messages into CAN bus. These malicious messages can directly intervene the functions inside AVs. Therefore, this article proposes a novel cross-chain-based intrusion detection for CAN bus (CCID-CAN) model that uses rule-based Valid Bit index (VBIN) model for initial intrusion detection on CAN bus inside AVs, followed by the Kalman filter and Naïve Bayes model for detecting attacks missed in the VBIN, where a cross-chain mechanism implements the exchange of attack logs among connected AVs that may not trust each other so as to optimize the Naïve Bayes detector. Afterwards, a series of experiments against several types of attacks are conducted on real vehicle supported by XPeng, and the results reveal that the CCID-CAN model outperforms existing models in terms of detection performance, time overhead, and memory footprint. In addition, attack log exchange for cross-chain networks in this proposed model is of high performance in latency, memory footprint, and throughput.

As a privacy-preserving machine learning paradigm, federated learning (FL) has attracted widespread attention from both academia and industry. Decentralized FL (DFL) overcomes the problems of untrusted aggregation server, single point of failure and poor scalability in traditional FL, making it suitable for industrial Internet of Things (IIoT). However, DFL provides more convenient conditions for malicious participants to launch attacks. This article focuses on the model poisoning attack in DFL for the first time, and proposes a novel attack method called collusive model poisoning attack (CMPA). To implement CMPA, we propose the dynamic adaptive construction mechanism, in which malicious participants can dynamically and adaptively construct malicious local models that meet distance constraints, reducing the convergence speed and accuracy of consensus models. Furthermore, we design the collusion-based attack enhancement strategies, where multiple participants can collude in the process of constructing malicious local models to improve the strength of attack. Empirical experiments conducted on MNIST and CIFAR-10 datasets reveal that CMPA significantly impacts the training process and results of DFL. Attack tests against representative defense methods show that CMPA not only invalidates statistical-based defenses but also skillfully overcomes performance-based methods, further proving its effectiveness and stealthiness. In addition, experiments based on practical IIoT scenario have also shown that CMPA can effectively disrupt system functionality.

Ethical understanding aims at morally analyzing and discriminating ethical scenarios described in natural language. By classifying behaviors that occur in ethical scenarios as ethical or unethical, ethical understanding empowers artificial intelligence systems to understand human values so as to discern right from wrong morally. However, most existing ethical understanding methods lack fine-grained analysis and cannot handle the problem that an ethical scenario may contain multiple behaviors with multiple polarities. In this paper, we introduce a novel natural language processing task, behavior-based ethical understanding (BEU), for mining the purpose relation(s) and ethical polarity of a specific behavior from the social news. It contains three subtasks: behavior term extraction (BTE) to extracts behavior terms, purpose relation inference (PRI) to identifies purposive relations among behaviors, and polarity discrimination (PD) to predicts the ethical polarities of behaviors, respectively. To perform this task, we constructed a Chinese BEU dataset, named FG-ETHICS. Besides, we propose a three-stage framework, BEU-BERT, based on the pre-trained language model BERT and deliberately designed downstream models for three subtasks. Experimental results show that the proposed framework achieves the best performance from the BTE and PD tasks, and achieves a promising performance of 75% on the PRI task.

Correlated high-utility pattern mining (CoUPM) considers the correlation between items in a pattern and offers a more reliable analysis for users. In real-world applications, the discovered patterns from CoUPM can present more interpretable information, but not all of them are useful. Generally, users pay attention to the number of items a pattern contains, which allows them to make reasonable decisions. In this paper, we solve the problem of mining those correlated high-utility patterns whose length is specified. A utility-list-based algorithm called Flexible Correlated Utility-based Pattern (FCoUP) is proposed. Furthermore, we propose some pruning strategies with the designed upper bounds for the two evaluation metrics: correlation and utility, reducing unwanted patterns generated and nodes visited during the mining process. Experiments show that FCoUP variants can produce more intelligent and flexible correlated high-utility patterns on a variety of datasets.

Abstract In context-aware recommendation systems, most existing methods encode users’ preferences by mapping item and category information into the same space, which is just a stack of information. The item and category information contained in the interaction behaviours is not fully utilized. Moreover, since users’ preferences for a candidate item are influenced by the changes in temporal and historical behaviours, it is unreasonable to predict correlations between users and candidates by using users’ fixed features. A fine-grained and coarse-grained information based framework proposed in our paper which considers multi-granularity information of users’ historical behaviours. First, a parallel structure is provided to mine users’ preference information under different granularities. Then, self-attention and attention mechanisms are used to capture the dynamic preferences. Experiment results on two publicly available datasets show that our framework outperforms state-of-the-art methods across the calculated evaluation metrics.

For blockchain software systems, framework developers may introduce technical debts that application developers are not aware of. Because these technical debts can have a negative impact on software projects, we need to investigate the issue of technical debt in blockchain software systems. We wanted to investigate what types of self-introduced technical debt exist in open-source blockchain software systems, and how these technical debts are distributed. We have selected six most popular blockchain software projects from GitHub. Then the code comments from these software projects were extracted and manually labelled. Finally, the code comments were statistically analysed. We propose a new type of technical debt, resource debt, which is explicitly identified by the framework developers and requires special attention in subsequent production systems. Six types of technical debt are prevalent and there is not any algorithm debt. In addition, we find that the code comments containing technical debt are not entirely determined by task tags. SATD is prevalent in blockchain projects. There is more significant variability between different application software projects for different technical debts. The results of the study imply that for detecting SATD, deep semantic discovery models should be used, such as pre-trained models.

Co-location pattern mining plays an important role in spatial data mining. With the rapid growth of spatial datasets, the usefulness of co-location patterns is strongly limited by the huge amount of discovered patterns. Although several methods have been proposed to reduce the number of discovered patterns, these statistical algorithms are unable to guarantee that the extracted co-location patterns are user preferred. Therefore, it is crucial to help the decision maker discover his/her preferred co-location patterns via efficient interactive procedures. This article proposes a new interactive approach that enables the user to discover his/her preferred co-location patterns. First, we present a novel and flexible interactive framework to assist the user in discovering his/her preferred co-location patterns. Second, we propose using ontologies to measure the similarity of two co-location patterns. Furthermore, we design a pruning scheme by introducing a pattern filtering model for expressing the user's preference, to reduce the number of the final output. By applying our proposed approach over voluminous sets of co-location patterns, we show that the number of filtered co-location patterns is reduced to several dozen or less and, on average, 80% of the selected co-location patterns are user preferred.

The methodology, formulating a reasonable task assignment to find the most suitable workers for a task and achieving the desired objectives, is the most fundamental challenge in spatial crowdsourcing. Many task assignment approaches have been proposed to improve the quality of crowdsourcing results and the number of task assignment and to limit the budget and the travel cost. However, these approaches have two shortcomings: (1) these approaches are commonly based on the attributes influencing the result of task assignment. However, different tasks may have different preferences for individual attributes; (2) the performance and efficiency of these approaches are expected to be improved further. To address the above issues, we proposed a task assignment approach in spatial crowdsourcing based on multiattribute decision-making (TASC-MADM), with the dual objectives of improving the performance as well as the efficiency. Specifically, the proposed approach jointly considers the attributes on the quality of the worker and the distance between the worker and the task, as well as the influence differences caused by the task’s attribute preference. Furthermore, it can be extended flexibly to scenarios with more attributes. We tested the proposed approach in a real-world dataset and a synthetic dataset. The proposed TASC-MADM approach was compared with the RB-TPSC and the Budget-TASC algorithm using the real dataset and the synthetic dataset; the TASC-MADM approach yields better performance than the other two algorithms in the task assignment rate and the CPU cost.

The widespread use of Internet of Things (IoT) technology has promoted location‐based service (LBS) applications. Users can enjoy various conveniences brought by LBS by providing location information to LBS. However, it also brings potential privacy threats to location information. Location data that contains private information is often transmitted among IoT networks in LBS, and such privacy information should be protected. In order to solve the problem of location privacy leakage in LBS, a location privacy protection scheme based on k‐anonymity is proposed in this paper, in which the Geohash coding model and Voronoi graph are used as grid division principles. We adopt the client‐server‐to‐user (CS2U) model to protect the user’s location data on the client side and the server side, respectively. On the client side, the Geohash algorithm is proposed, which converts the user’s location coordinates into a Geohash code of the corresponding length. On the server side, the Geohash code generated by the user is inserted into the prefix tree, the prefix tree is used to find the nearest neighbors according to the characteristics of the coded similar prefixes, and the Voronoi diagram is used to divide the area units to complete the pruning. Then, using the Geohash coding model and the Voronoi diagram grid division principle, the G‐V anonymity algorithm is proposed to find k neighbors in an anonymous area so that the user’s location data meets the k‐anonymity requirement in the area unit, thereby achieving anonymity protection of location privacy. Theoretical analysis and experimental results show that our method is effective in terms of privacy and data quality while reducing the time of data anonymity.

With the development of emerging intelligent traffic signal (I-SIG) system, congestion-involved security issues are drawing attentions of researchers and developers on the vulnerability introduced by connected vehicle technology, which empowers vehicles to communicate with the surrounding environment such as road-side infrastructure and traffic control units. A congestion attack to the controlled optimization of phases algorithm (COP) of I-SIG is recently revealed. Unfortunately, such analysis still lacks a timely visualized prediction on later congestion when launching an initial attack. In this paper, we argue that traffic image feature-based learning has available knowledge to reflect the relation between attack and caused congestion and propose a novel analysis framework based on cycle generative adversarial network (CycleGAN). Based on phase order, we first extract four-direction road images of one intersection and perform phase-based composition for generating new sample image of training. We then design a weighted L1 regularization loss that considers both last-vehicle attack and first-vehicle attack, to improve the training of CycleGAN with two generators and two discriminators. Experiments on simulated traffic flow data from VISSIM platform show the effectiveness of our approach.