This systematic review examines harassment in virtual reality (VR), synthesizing findings from 85 studies published between 2017 and 2025. We propose a nuanced typology of harassment, encompassing spatial intrusion, sexual and verbal abuse, identity-based discrimination, group-targeted harassment, and systemic harms, and demonstrate how VR’s immersive and embodied features amplify risk and impact. Marginalized users, such as women, LGBTQ+ individuals, children, and people with disabilities, face disproportionate harm. We further analyze the psychological and behavioral consequences of harassment, as well as the effectiveness and limitations of current governance, design, and AI-driven interventions. Our review identifies persistent research gaps in theory, measurement, and inclusive protection, and advocates for ethical, participatory, and preventive approaches to platform safety. This work aims to guide researchers and designers in building more equitable and safe VR environments.

Abstract Although packers are useful tools for protecting applications, they can also be used to protect malware. This makes packer identification technology increasingly important for malware analysis. Most existing packer identification methods based on static analysis extract a large number of features from the binary code of an executable. However, these binary code-related features are sensitive to small changes in the executable’s binary code, and it is difficult to understand how these features influence the decision-making processes of machine learning models. To address the shortcomings of existing static analysis-based packer identification methods, we explore extraction of a small number of easily extractable and discriminative features for efficient and accurate packer identification. Specifically, we analyze the packing process and notice that the structures of the packed PEs differ according to the different packing patterns used by the packer. Based on this, a section-entropy plot is proposed, which is generated by a small number of easily extractable and discriminative features that can reflect the overall structure of a PE file. By using GoogLeNet to identify the packer characterized by the section-entropy plot, a p acker i dentification method based on the s ection- e ntropy p lot (PISEP) is constructed, which does not require PE file disassembly and complex feature engineering. The experimental results show that PISEP achieves 99.08% accuracy for identifying seen and unseen types of packers and requires only 0.165 s on average to identify the packer class of a test sample, and thus could be a highly competitive candidate for packer identification.

Abstract In response to the increasing threat posed by the exponential growth of malware in cybersecurity, researchers have developed a number of malware classification methods based on malware images and deep learning in recent years. Newly proposed methods of this type tend to focus on generating malware images by extracting multiple types of information from a PE file, as well as on using complex convolutional neural network (CNN) models, to achieve high classification accuracy. Methods that involve extracting multiple types of information, especially those that require file disassembly for acquisition and the subsequent use of complex CNN models, result in a lengthy process for generating malware images and significantly increase model training durations. To alleviate this problem, we adopt the idea of using only a small part of the content that can be easily extracted from a PE file to efficiently generate a malware image, and implement malware classification without relying on complex CNN models. As a key component of a PE file, the PE header and the section table (we call them PE metadata) are characterized by a relatively low byte count and are likely to be useful for malware classification according to the similarities observed in the PE metadata between malware from both the same family and different families. Therefore, in this work, we explore the feasibility of using PE metadata alone to generate an image for malware classification and propose an Image of PE metadata (IPM) generated from PE metadata to represent malware. Based on the proposed IPM, we then construct a shallow CNN model and combine it with a support vector machine classifier to introduce a novel malware classification method called MCPDS ( M alware c lassification method using P E metadata, d eep learning and s upport vector machine). The experimental results show that the MCPDS not only achieves high accuracy in terms of classifying malware on two malware datasets but also exhibits high efficiency in terms of image generation and good robustness against adversarial samples.

Federated learning offers a partial safeguard for participants’ data privacy. Nevertheless, the current absence of an efficient privacy-preserving federated learning technology tailored for the Internet of Things (IoT) poses a challenge. Numerous privacy-preserving federated learning frameworks have been proposed, primarily relying on homomorphic cryptosystems, yet their suitability for IoT remains limited. Furthermore, the application of federated learning in IoT confronts two significant obstacles: mitigating the substantial communication costs and communication failure rates, and effectively discerning and utilizing high-quality data while discarding low-quality data for collaborative modeling purposes. In order to address these challenges, this paper introduces a privacy-preserving optimal aggregation federated learning framework that relies on the utilization of the multi-key EC-ElGamal cryptosystem (MEEC) and the federated sum optimization algorithm (FSOA), which are characterized by their lightweight nature and fair properties. The proposed MEEC approach aims to tackle the issue of multi-key collaborative computing within the context of federated learning, thereby resulting in reduced communication costs and enhanced communication efficiency. This is achieved through the leverage of the EC-ElGamal cryptosystem, which is known for its ability to generate short keys and ciphertexts. Furthermore, this paper presents a dynamic federated learning framework that incorporates user dynamic quit and join algorithms. The primary objective of this framework is to mitigate the adverse effects of communication failures and enhance power computation on IoT devices. Additionally, an FSOA is devised to ensure the acquisition of optimal training data, thereby preventing the inclusion of low-quality data in the training process. Subsequently, the proposed scheme undergoes rigorous security analysis and performance evaluation. The obtained results unequivocally demonstrate that our scheme outperforms existing solutions in terms of security, practicality, and efficiency with lower communication and computational costs.

With the rapid development of the internet worldwide, network security issues are becoming increasingly prominent. Network intrusion detection systems (NIDSs) play a vital role in ensuring computer network security due to their ability to identify potential network threats. Despite considerable research efforts, deploying NIDSs on resource-constrained devices has been challenging. To reduce the imposed computational cost and model storage requirements, in this paper, we propose a novel lightweight NIDS model. In this model, patterns of normal and malicious actions are learned via a finite Dirichlet mixture model (DMM) in the context of the extended stochastic variational inference (ESVI) framework. With the proposed method, both the parameter estimation and model selection processes can be simultaneously addressed in a unified Bayesian framework. A great number of experiments conducted on three publicly available datasets demonstrate that the proposed model not only achieves comparable classification performance to that of detection models based on several well-studied finite mixture modeling, traditional machine learning (ML) and promising deep learning (DL) algorithms but also significantly reduces the required training and detection time. Extensive experimental results validate that the proposed model is a feasible and efficient lightweight intrusion detection model.

Bayesian estimation of parameters in the Dirichlet mixture process of the Beta-Liouville distribution (i.e., the infinite Beta-Liouville mixture model) has recently gained considerable attention due to its modeling capability for proportional data. However, applying the conventional variational inference (VI) framework cannot derive an analytically tractable solution since the variational objective function cannot be explicitly calculated. In this paper, we adopt the recently proposed extended VI framework to derive the closed-form solution by further lower bounding the original variational objective function in the VI framework. This method is capable of simultaneously determining the model's complexity and estimating the model's parameters. Moreover, due to the nature of Bayesian nonparametric approaches, it can also avoid the problems of underfitting and overfitting. Extensive experiments were conducted on both synthetic and real data, generated from two real-world challenging applications, namely, object detection and text categorization, and its superior performance and effectiveness of the proposed method have been demonstrated.

Industrial Internet of Things (IIoT) is gradually changing the mode of traditional industries with the rapid development of big data. Besides, thanks to the development of deep learning, it can be used to extract useful knowledge from the large amount of data in the IIoT to help improve production and service quality. However, the lack of large-scale data sets will lead to low performance and overfitting of learning models. Therefore, federated deep learning with distributed data sets has been proposed. Nevertheless, the research has shown that federated learning can also leak the private data of participants. In IIoT, once the privacy of participants in some special application scenarios is leaked, it will directly affect national security and people’s lives, such as smart power grid and smart medical care. At present, several privacy-preserving federated learning schemes have been proposed to preserve data privacy of participants, but security issues prevent them from being fully applied. In this article, we analyze the security of the DeepPAR scheme proposed by Zhang et al., and point out that the scheme is insecure in the re-encryption key generation process, which will cause the leakage of the secret key of participants or the proxy server. In addition, the scheme is not resistant to collusion attacks between the parameter server and participants. Based on this, we propose an improved scheme. The security proof shows that the improved scheme solves the security problem of the original scheme and is resistant to collusion attacks. Finally, the security and accuracy of the scheme is illustrated by performance analysis.

Abstract Format‐preserving encryption (FPE) allows encrypting plaintexts while preserving a specific format. In Selected Areas in Cryptography 2018, two targeted ciphers were proposed as new FPE schemes. The second scheme was designed with an algorithm called Mix–Swap–Unmix that is shown to be equivalent to a particular matching exchange process under a specific setting. In this comment paper, we prove that the matching exchange process is invalid. As a result, this equivalence does not exist.

Due to the transparency of the wireless channel, users in multiple-key environment are vulnerable to eavesdropping during the process of uploading personal data and re-encryption keys. Besides, there is additional burden of key management arising from multiple keys of users. In addition, profile matching using inner product between vectors cannot effectively filter out users with ulterior motives. To tackle the above challenges, we first improve a homomorphic re-encryption system (HRES) to support a single homomorphic multiplication and arbitrarily many homomorphic additions. The public key negotiated by the clouds is used to encrypt the users’ data, thereby avoiding the issues of key leakage and key management, and the privacy of users’ data is also protected. Furthermore, our scheme utilizes the homomorphic multiplication property of the improved HRES algorithm to compute the cosine result between the normalized vectors as the standard for measuring the users’ proximity. Thus, we can effectively improve the social experience of users.

Data clustering is the unsupervised classification of data records into groups. As one of the steps in data analysis, it has been widely researched and applied in practical life, such as pattern recognition, image processing, information retrieval, geography, and marketing. In addition, the rapid increase of data volume in recent years poses a huge challenge for resource-constrained data owners to perform computation on their data. This leads to a trend that users authorize the cloud to perform computation on stored data, such as keyword search, equality test, and outsourced data clustering. In outsourced data clustering, the cloud classifies users’ data into groups according to their similarities. Considering the sensitive information in outsourced data and multiple data owners in practical application, it is necessary to develop a privacy-preserving outsourced clustering scheme under multiple keys. Recently, Rong et al. proposed a privacy-preserving outsourced k-means clustering scheme under multiple keys. However, in their scheme, the assistant server (AS) is able to extract the ratio of two underlying data records, and key management server (KMS) can decrypt the ciphertexts of owners’ data records, which break the privacy security. AS can even reduce all data records if it knows one of the data records. To solve the aforementioned problem, we propose a highly secure privacy-preserving outsourced k-means clustering scheme under multiple keys in cloud computing. In this paper, noncolluded cloud computing service (CCS) and KMS jointly perform clustering over the encrypted data records without exposing data privacy. Specifically, we use BCP encryption which has additive homomorphic property and AES encryption to double encrypt data records, where the former cryptosystem prevents CCS from obtaining any useful information from received ciphertexts and the latter one protects data records from being decrypted by KMS. We first define five protocols to realize different functions and then present our scheme based on these protocols. Finally, we give the security and performance analyses which show that our scheme is comparable with the existing schemes on functionality and security.

Homomorphic encryption is widely used in the scenarios of big data and cloud computing for supporting calculations on ciphertexts without leaking plaintexts. Recently, Li et al. designed a symmetric homomorphic encryption scheme for outsourced databases. Wang et al. proposed a successful key-recovery attack on the homomorphic encryption scheme but required the adversary to know some plaintext/ciphertext pairs. In this paper, we propose a new ciphertext-only attack on the symmetric fully homomorphic encryption scheme. Our attack improves the previous Wang et al.’s attack by eliminating the assumption of known plaintext/ciphertext pairs. We show that the secret key of the user can be recovered by running lattice reduction algorithms twice. Experiments show that the attack successfully and efficiently recovers the secret key of the randomly generated instances with an overwhelming probability.

The knapsack problem is an important problem in computer science and had been used to design public key cryptosystems. Low‐density subset sum algorithms are powerful tools to reduce the security of trapdoor knapsacks to the shortest vector problem (SVP) over lattices. Several knapsack ciphers Chor–Rivest, Okamoto–Tanaka–Uchiyama, and Kate–Goldberg were proposed to defend low‐density attacks by utilising low‐weight knapsack problems. Some evidence was also found on the vulnerabilities of the above three knapsack ciphers to lattice attacks. However, previous lattice‐based cryptanalytic results have been established via a probabilistic approach. The authors investigate some collision‐free properties and derive from the properties a deterministic reduction from the knapsack problems in the Chor–Rivest, Okamoto–Tanaka–Uchiyama, and Kate–Goldberg knapsack ciphers to SVP without imposing any restriction and assumption. To the best of the authors' knowledge, the proposed reduction is the first deterministic reduction from public key cryptographic knapsacks to SVP.

Multivariate public key cryptography is a set of cryptographic schemes built from the NP-hardness of solving quadratic equations over finite fields, amongst which the hidden field equations (HFE) family of schemes remain the most famous. However, the original HFE scheme was insecure, and the follow-up modifications were shown to be still vulnerable to attacks. In this paper, we propose a new variant of the HFE scheme by considering the special equation x2=x defined over the finite field F3 when x=0,1 . We observe that the equation can be used to further destroy the special structure of the underlying central map of the HFE scheme. It is shown that the proposed public key encryption scheme is secure against known attacks including the MinRank attack, the algebraic attacks, and the linearization equations attacks. The proposal gains some advantages over the original HFE scheme with respect to the encryption speed and public key size.

As an important boundary-based clustering algorithm, support vector clustering (SVC) can benefit many real applications owing to its capability of handling arbitrary cluster shapes, especially those directly or indirectly related to pattern exploration and description. As the application deepens, the importance of performance (i.e. criterions of accuracy and efficiency) of SVC increases. To identify gaps in the current methods and propose novel research directions for SVC, we present a survey of the literature in this area. Our approach is to classify the most recent advances into either theory or application. For theoretical contributions, advances related to parameter selection and optimization, dual-problem solutions, and cluster labeling are introduced. We also simultaneously summarize the advantages and drawbacks of each study. With respect to applications, we clearly describe eight groups of schemes based on SVC, either as individual or hybrid methods. Finally, we identify the gaps in SVC research and suggest several future research issues and trends.