The industrial internet of things (IIoT) has been widely applied in fields such as smart manufacturing, energy management, and infrastructure monitoring. Its distributed nature and cross-domain data collaboration requirements make federated learning (FL) an important modeling paradigm. However, FL systems in IIoT are vulnerable to backdoor attacks, where adversaries implant hidden trigger patterns during local model training to induce the global model to produce malicious outputs for specific inputs. To address this issue, we propose a federated backdoor defense method named FedQmc, which is based on quantum state encoding and multi-scale aggregation. Specifically, unlike traditional detection methods that rely on geometric distances in parameter space, we design a quantum state encoding-driven anomaly measurement mechanism. By encoding local updates into quantum states, model updates can be represented in high-dimensional complex Hilbert space, and the coherence and superposition properties of quantum states are exploited to enhance feature discrimination, thereby providing significant advantages in capturing subtle yet highly correlated backdoor perturbations. Based on this, we propose multi-scale robust aggregation, which simultaneously suppresses different types of backdoor updates at three levels: layer scale, block scale, and low-rank subspace scale. In addition, we develop a meta-learning-based adaptive weight optimization mechanism, which enables the aggregation weights at each scale to be dynamically adjusted across multiple training rounds. This allows FedQmc to resist stronger and stealthier backdoor attacks while maximally retaining useful gradient information from benign clients. Theoretical analysis demonstrates the convergence and security guarantees of FedQmc. Experimental results show that FedQmc significantly reduces the attack success rate while maintaining high accuracy across multiple datasets, thereby effectively enhancing multi-scale defense capability in IIoT scenarios.

Online reviews on platforms such as Amazon and Yelp significantly influence consumer decisions and business reputations. However, spammers form groups that strategically control product reviews over specific periods to manipulate consumer sentiment and decision-making. Traditional spammer group detection methods face two primary issues: 1) knowledge marginalization: interactions dominate the model to form candidate groups, potentially marginalizing valuable structured knowledge; and 2) temporal knowledge discrepancy: inconsistencies in users’ temporal activities and behavioral features lead to blurry classification of candidate groups. To address these two limitations, we introducetemporal knowledge consistency for spammer groups detection via contrastive learningcalled TKCCL. We employ dual-view encoders derived from knowledge graphs and heterogeneous information networks to learn informative representations, thereby alleviating knowledge marginalization. TKCCL maps the temporal knowledge as vectors to measure consistency, enhancing users’ rating proximity and temporal synchronization in dual views, thereby reducing temporal knowledge discrepancies. We optimize spammer group detection by modeling it as a greedy set cover problem, which enhances the method’s responsiveness to dynamic spam strategies. Experimental results on four public datasets demonstrate that TKCCL substantially outperforms existing methods. Our code is available athttps://github.com/NeenLee/TKCCL.

Abstract For semantic segmentation of cardiac magnetic resonance image (MRI) with low recognition and high background noise, a fusion‐attention Swin Transformer is proposed based on cognitive science and deep learning methods. It has a U‐shaped symmetric encoding–decoding structure with an attention‐based skip connection. The encoder realizes self‐attention for deep feature representation and the decoder up‐samples global features to the corresponding input resolution for pixel‐level segmentation. By introducing a skip connection between the encoder and decoder based on fusion attention, the remote interaction of global information is realized, and the attention to local features and specific channels is enhanced. A public ACDC cardiac MRI image dataset is used for experiments. The segmentation of the left ventricle, right ventricle, and myocardial layer is realized. The method performs well on a small sample dataset, for example, the pixel accuracy obtained by the proposed model is 93.68%, the Dice coefficient is 92.28%, and HD coefficient is 11.18. Compared with the state‐of‐the‐art models, the segmentation precision has been significantly improved, especially for the low recognition and heavily occluded targets.

The blockchain-enabled cloud manufacturing is an emerging service-oriented paradigm, and the scheduling and process optimization for blockchain-enabled cloud manufacturing (SPO-BCMfg) are crucial to achieving the service-oriented goal. The blockchain-enabled cloud manufacturing paradigm improves the collaboration capabilities and information security over the ordinary cloud manufacturing while incorporating distributed storage, consensus mechanism, and cloud-edge collaboration. The above characteristics make SPO-BCMfg a multiobjective scheduling optimization problem. This article establishes the multiobjective SPO-BCMfg model based on a dynamic selection evolutionary algorithm to address the problem. First, we carry out the architecture and the modeling of the blockchain cloud manufacturing system. Then, a novel dynamic selection evolutionary algorithm is proposed, which is used to schedule and optimize the model for the process. In the stage of evolution, the algorithm uses a diversity-based population partitioning technique that utilizes the dynamic distance to realize the selection of elite solutions. The method was experimented on the SPO-BCMfg problem facing five and eight objectives. The experimental results show that the algorithm has a strong processing capacity in terms of convergence and diversity compared with the other advanced evolutionary algorithms.

Aiming to address the unsatisfactory performance of existing distributed deep learning architectures, such as poor accuracy, slow network communication, low arithmetic speed, and insufficient security, we propose and design a learning model based on a distributed deep learning and blockchain architecture. We use a hybrid parallel algorithm based on blockchain (HP-B) to build a distributed deep consensus learning model. The HP-B algorithm is grouped according to the performance of computing nodes participating in training, network links and training samples, and the grouped computing equipment performs optimal distributed computing. The purpose of this approach is to solve the security and scalability concerns and improve the convergence speed and accuracy of deep learning. The proposed method achieves good results on the CIFAR-100, CIFAR-10, and IMAGENET data sets. Finally, the distributed deep learning model based on blockchain is combined with the generative adversarial network to solve the segmentation problem of medical imaging data, and the experimental results are superior to those of other networks.

With rapid growth in data volume generated from different industrial devices in IoT, the protection for sensitive and private data in data sharing has become crucial. At present, federated learning for data security has arisen, and it can solve the security concerns on data sharing by model sharing on Internet of mutual distrust. However, the hackers still launch attack aiming at the security vulnerabilities (e.g., model extraction attack and model reverse attack) in federated learning. In this article, to address the above problems, we first design an application model of blockchain-enabled federated learning in Industrial Internet of Things (IIoT), and formulate our data protection aggregation scheme based on the above model. Then, we give the distributed K-means clustering based on differential privacy and homomorphic encryption, and the distributed random forest with differential privacy and the distributed AdaBoost with homomorphic encryption methods, which enable multiple data protection in data sharing and model sharing. Finally, we integrate the methods with blockchain and federated learning, and provide the complete security analysis. Extensive experimental results show that our aggregation scheme and working mechanism have the better performance in the selected indicators.

Due to the flexibility of heterogeneous information networks (HINs) in modeling heterogeneous data, researchers begin to use it to integrate the objects and relationships in recommender systems. However, how to abstract and exploit effective information and apply the information to recommender systems is a challenge. To fully mine nodes' structural features and better integrate these features simultaneously, we present a nonlinear feature fusion-based rating prediction algorithm. This algorithm first uses a meta-path-based HIN embedding model to extract the nodes' structural features. Then, the structural features are converted by a nonlinear fusion method. Finally, the fused features are input into the multilayer perceptron to achieve rating prediction. Experiments on real-life data sets, such as Movielens-100k, Yelp, Douban Book, and Douban Movie, are designed to prove the performance of the proposed model. Experimental results on the four data sets reveal that our algorithm is superior to the baselines.

The article A whitelist and blacklist-based co-evolutionary strategy for defensing against multifarious trust attacks, written by Shujuan Ji, Haiyan Ma, Yongquan Liang, Hofung Leung and Chunjin Zhang, was originally published electronically on the publisher’s internet portal.

Due to uneven deployment of anchor nodes in large‐scale wireless sensor networks, localization performance is seriously affected by two problems. The first is that some unknown nodes lack enough noncollinear neighbouring anchors to localize themselves accurately. The second is that some unknown nodes have many neighbouring anchors to bring great computing burden during localization. This paper proposes a localization algorithm which combined niching particle swarm optimization and reliable reference node selection in order to solve these problems. For the first problem, the proposed algorithm selects the most reliable neighbouring localized nodes as the reference in localization and using niching idea to cope with localization ambiguity problem resulting from collinear anchors. For the second problem, the algorithm utilizes three criteria to choose a minimum set of reliable neighbouring anchors to localize an unknown node. Three criteria are given to choose reliable neighbouring anchors or localized nodes when localizing an unknown node, including distance, angle, and localization precision. The proposed algorithm has been compared with some existing range‐based and distributed algorithms, and the results show that the proposed algorithm achieves higher localization accuracy with less time complexity than the current PSO‐based localization algorithms and performs well for wireless sensor networks with coverage holes.

Underwater sensor networks have many applications ranging from ocean monitoring, undersea exploration, target tracking, coastal surveillance, to disaster prevention. In multi-application scenarios, the network might need to handle different types of packets to satisfy the requirements of diverse data transmission metric. For example, multimedia-based applications may include different multimedia packets, such as voice, compressed images, even video streams with different quality of experience. To meet the requirements of such applications, in this paper, we propose a hybrid adaptive routing scheme (HARS) for drifting restricted floating ocean sensor networks (DR-OSNs), which exploits both surface wireless and underwater communication channels to fulfill different performance requirements. We evaluate the performance of the routing scheme and investigate the factors which affect the scheme. The simulation results demonstrate that the scheme achieves a reasonable performance for different communication channels and packet delivery.

With electronic commerce becoming increasingly popular, the problems of trust have become one of the main challenges in the development of electronic commerce. Although various mechanisms have been adopted to guarantee trust between customers and sellers (or platforms), trust and reputation systems are still frequently attacked by deceptive, collusive, or strategic agents. Therefore, it is difficult to keep these systems robust. It has been mentioned that a combined usage of both trust and distrust propagation can lead to better results. However, little work has been known to realize this insight successfully. Besides, literatures either use a social network with trust/distrust information or use one advisor list in evaluating all sellers, which leads to the lack of pertinence and inaccuracy of evaluation. This paper proposes a defensing strategy called WBCEA , in which, each buyer agent is modeled with two attributes (i.e., the trustworthy facet and the untrustworthy facet) and two lists (i.e., the whitelist and the blacklist). Based on the social network that are constructed and maintained according to its whitelist and blacklist, the honest buyer agent can find trustable buyers and evaluate the candidate sellers according to its own experience and ratings of trustable buyers. Experiments are designed and implemented to verify the accuracy and robustness of this strategy. Results show that our strategy outperforms existing ones, especially when majority of buyers are dishonest in the electronic market.