The rapid advancement of the Internet of Vehicles (IoV) necessitates robust solutions that balance user privacy with efficient service delivery, particularly for personalized charging station recommendations. This paper proposes a lightweight anonymous authentication and secure computing service integration scheme based on distributed center architecture. By leveraging a distributed dual-cloud architecture, the scheme ensures secure and traceable anonymous authentication while facilitating efficient vehicle-to-service communication. Vehicle users encrypt their preference data, which is processed using matrix factorization and secure multi-party computation to deliver accurate, privacy-preserving charging station recommendations. The quick reconnection mechanism we established reduces re-authentication time costs by roughly 53 percent, while the dual-server architecture supports privacy computing for vehicles after authentication. The framework integrates multi-source heterogeneous data, such as real-time traffic and charging station status, to enhance recommendation accuracy and timeliness. Experimental results demonstrate that the proposed approach achieves strong anonymity, traceability, and low-latency performance, offering a practical and scalable solution for secure, intelligent IoV services.

Skyline query is a fundamental technique in multi-criteria decision-making, aiming to extract "optimal" results that are not dominated by any other data points across all attributes. It has significant value in applications that require trade-offs among multiple criteria. However, existing skyline query methods face two critical limitations: (i) conventional approaches adopt fixed dominance relationships, making it difficult to capture personalized user preferences; and (ii) cloud-based deployment models risk exposing sensitive data and query logic, making it difficult to ensure data privacy and protect query patterns while maintaining efficiency. To address these issues, we propose Privacy-Preserving User-Defined Skyline Query (PUDSQ), a novel privacy-preserving user-defined skyline query framework, which integrates efficient cryptographic techniques-secret sharing (SS) and function secret sharing (FSS)-with a secure database shuffling mechanism to achieve efficient query processing while ensuring robust privacy guarantees. PUDSQ introduces three main innovations: (i) a privacy-preserving filtering framework based on FSS provides dual protection for both data content and user preferences, effectively concealing database content and query logic; (ii) an FSS-based secure protocol suite supporting user-defined attribute retrieval, constrained-region retrieval, and secure skyline filtering; and (iii) a high-dimensional data processing strategy that integrates dimensionality reduction with an Sort-Filter-Skyline (SFS)-based presorting approach to address the high-dimensional data processing challenge and significantly improve efficiency. Experimental results demonstrate that, under equivalent security guarantees, PUDSQ reduces query latency by 8%-90% compared with state-of-the-art solution, with particularly notable advantages in high-dimensional scenarios, achieving an effective efficiency-privacy trade-off.

Federated learning (FL), as a distributed machine learning paradigm, has gained widespread adoption due to its ability to retain user data locally, thereby protecting privacy, while collaboratively training a global model through gradient sharing. However, existing studies have shown that attackers may obtain privacy information from the gradients, and malicious server may return erroneous aggregated results, compromising federated learning model. Although prior studies have addressed privacy preservation and aggregated result verification, these methods often incur significant computation and communication overhead on the user side. In this paper, we propose an efficient privacy-preserving and verifiable federated learning (EPVFL) scheme. Specifically, we group the gradients and employ polynomial encryption to achieve efficient privacy protection. Furthermore, we design a lightweight verification mechanism where users only need to perform lightweight local computation without interaction and transmit just a floating-point vector to verify the correctness of the aggregated gradient. EPVFL supports users going offline at any time, while online users can still obtain the correct aggregated gradient without incurring additional computation or communication overhead. Finally, through security analysis and experiments on real datasets, we demonstrate the correctness, verifiability, and privacy protection of EPVFL. Experiment results indicate that EPVFL protects privacy without sacrificing model accuracy and significantly reduces the computation and communication overheads on the user side compared to the related schemes.

Searchable Encryption (SE) enables users to perform searches on encrypted data while preserving data privacy. Since cloud servers are platforms that provide services for a large number of users, and data owners require access control over their data, SE schemes that support multi-user settings and access control are therefore more suitable for cloud storage. However, in existing SE schemes that support multi-user settings and access control, most only support single-keyword or conjunctive keyword searches, and the search time grows linearly with the total amount of data. These limitations negatively impact both the accuracy and efficiency of search operations. This work proposes an SE scheme specifically designed for multi-user settings. Data owners can enforce fine-grained access control policies, while a specialized retrieval structure allows the cloud to assist users in performing Boolean keyword searches with improved efficiency. The search complexity of the proposed scheme is O(r) , where r denotes the number of files relevant to the queried keyword. We demonstrate the scheme's effectiveness and practicality through performance analysis.

With the increasing prevalence of outsourcing images to cloud servers, privacy-preserving content-based image retrieval (CBIR) has attracted significant research attention. Existing privacy-preserving CBIR schemes often prioritize retrieval speed by adopting methods that provide weak privacy guarantees and low-dimensional image features, which inevitably compromises security and retrieval accuracy. Additionally, most solutions directly employ CNN models pre-trained on public datasets for feature extraction, neglecting domain adaptation problem. To address these limitations, we propose SecOIR, a secure outsourced image retrieval scheme based on deep hashing networks, which achieves provable security under the semi-honest adversary model while hiding access patterns. Furthermore, domain adaptation is resolved through fine-tuning of feature extraction models. Experimental results demonstrate that SecOIR outperforms state-of-the-art schemes by 11%-12% in accuracy under identical datasets and configurations, while maintaining practical efficiency. To achieve SecOIR, we propose two modified function secret sharing (FSS) schemes that overcome the limited compatibility of the original FSS schemes with replicated secret sharing (RSS). Then, building upon the modified FSS schemes and RSS, we design a series of efficient sub-protocols. Benchmark tests reveal that our sub-protocols surpass existing mainstream solutions in efficiency, which can also serve as independent contributions to secure multi-party computation protocol design.

Private set intersection (PSI) faces significant challenges in handling dynamic data in the digital era. The data generated by high-frequency updates in real-world scenarios, such as real-time fraud detection and dynamic targeted advertising, has extremely high timeliness value, yet traditional PSI protocols designed for static data struggle to support realtime collaboration needs. While updatable PSI (UPSI) aims to address this issue, the state-of-the-art solutions predominantly rely on homomorphic encryption, whose public-key cryptography operations lead to inefficient online performance, unable to meet practical requirements for high-frequency incremental updates. To address these challenges, we propose a novel PSI protocol based on distributed multi-point functions. Our protocol breaks through the limitations of the traditional 2-server-1-client model in function secret sharing-based PSI schemes, thereby better aligning with practical two-party PSI application requirements. We achieve efficient online computation under the semi-honest model and further construct an UPSI scheme that avoids the performance overhead of public-key operations. Experiments demonstrate that our protocol surpasses the state-of-the-art homomorphic encryption-based scheme in online efficiency, achieving a speedup of 2.5-165.4 & times;. In a 30-day dynamic data simulation, our solution maintained only 7%-20% of the baseline's cumulative time cost, validating its feasibility for high-frequency update scenarios and providing a viable technical pathway for scenarios relying on updatable data collaboration.

Federated learning enables training without uploading raw data by exchanging model updates, yet gradients can still leak private information and malicious clients can poison training. Prior work typically protects privacy or detects malicious updates-or supports both at the cost of heavy communication. Communication is often the bottleneck in practical FL deployments. We propose Cerflan, a communication-efficient norm-bounding framework that privately screens updates under a dual-server, semi-honest model. Cerflan introduces a norm-aware masking scheme and two lightweight verification sub-protocols to enable masked 8 infinity and 82 checks and aggregation without oblivious transfer or Beaver triples. This trades 10-30 % extra computation for 40-60 % less communication than ELSA, making Cerflan attractive under limited bandwidth. In experiments, Cerflan curbs fast gradient sign method and BadNets attacks while preserving model utility.

The rapid growth of sensitive cross-domain data, such as electronic health records and genomic sequences in healthcare, presents significant opportunities for large-scale, multi-institutional collaborative analysis. Meeting stringent privacy regulations while utilizing data has become a critical challenge. Private set operations (PSO) play a crucial role to address this challenge. PSO protocols enable privacy-preserving data alignment across parties (such as interinstitutional data matching based on private set intersection (PSI)) and secure data aggregation (such as federated data aggregation through private set union (PSU)), providing fundamental support for cross-domain data collaboration. This type of technology is not only applicable to multi-center medical research but also has broad value in other scenarios requiring confidential data sharing. However, existing delegated/outsourced PSO schemes face two key limitations: 1) high client-side preprocessing overhead, requiring clients to expensively mask private data before uploading; 2) performance and single-point dependency bottlenecks in client-assisted computation where one client must act as a computational leader. To address these issues, we propose a secret-shared PSO framework for lightweight clients. In our scheme, the clients can go offline while servers perform all computations, significantly reducing clients' burden. Notably, our protocol can be extended to support multi-party settings, making it well-suited for collaborative research across multiple institutions. In addition, we prove the security of all constructions under the semi-honest model. Experiments show that when the set size n >= 2(16) , our protocol has a significant advantage and is well-suited for lightweight client that holds a large set.

Generating high-quality 3D point clouds requires capturing both global shape topology and local geometric details. Existing flow-based methods rely on continuous normalizing flows (CNFs) that demand expensive ODE solving and trace estimation during training, while diffusion models require hundreds of iterative denoising steps. Moreover, most approaches adopt single-level generation directly in point space, disregarding the hierarchical structure natural to 3D shapes. We propose Hierarchical Flow Matching (HFM) that extends flow matching to bilevel structure for unconditional 3D point cloud generation. HFM decomposes the task into two levels via optimal-transport flow matching: a Latent Flow Matching models the global shape manifold in a compact latent space, and a Conditional Point Flow Matching reconstructs detailed point clouds conditioned on the latent code. Both flows are trained with simple MSE regression losses. The resulting straight OT paths enable efficient sampling with as few as 15 Euler steps per flow, while the structured latent space supports downstream tasks including classification. Extensive experiments on ShapeNet and ModelNet benchmarks demonstrate that HFM achieves competitive or even best performance compared with prior state-of-the-art methods.

Abstract Extreme Gradient Boosting (XGBoost) is widely used due to its excellent learning performance. In real-world scenarios, data is often held by multiple owners, the sample size of a single data owner is usually insufficient for effectively training the model. Therefore, how to integrate data resources from different owners to collaboratively train a higher-performing model while ensuring data privacy has become an urgent problem that needs to be solved. To solve the above problems, we propose a privacy-preserving XGBoost scheme (PriXGB) with wide applicability and design a series of secure protocols for its key steps. We proposed a SecureArgmax protocol based on function secret sharing, which significantly improves the efficiency of securely computing the maximum value in the secret sequence while ensuring that intermediate information is not leaked. We designed a privacy sigmoid function protocol based on secret sharing to improve the accuracy of the model. In the vertical data partition scenario, PriXGB is roughly 30% more efficient than existing solutions. Finally, we provide security proofs for designed protocols and verify the advantages of the scheme through experiments. The experimental results show that PriXGB effectively protects data privacy while maintaining accuracy comparable to non-private XGBoost, thereby demonstrating strong security and effectiveness.

Data aggregation, as a data-processing technique, facilitates accurate diagnosis in the healthcare Internet of Things (HIoT) by integrating multisource heterogeneous health data. However, achieving efficient and secure aggregation of multidimensional medical data remains challenging, particularly when simultaneously preserving location privacy and providing fair, quality-driven incentives. To address these issues, this article proposes a quality-aware secure multidimensional data aggregation (QSDA) scheme with location privacy for HIoT. First, the scheme employs inner product encryption to support aggregation task matching without revealing users' actual coordinates and further integrates symmetric homomorphic encryption (SHE) with super-increasing sequences to enable one-stop compressed aggregation of multidimensional data, thereby effectively supporting common statistical operations such as mean and variance. Second, it introduces a data quality incentive mechanism based on offset metrics, while leveraging blockchain (BC) auditing to ensure the traceability of the aggregation process and the verifiability of the aggregation results. Finally, security analysis and performance evaluation demonstrate the scheme's effectiveness and efficiency.

The rapid growth of sensitive labeled data in healthcare, finance, user profiling, and commercial databases has created increasing demand for privacy-preserving collaborative data validation across different organizations. Following prior cryptographic studies, this paper uses private collaborative data cleaning (PCDC) to refer to a specific label-conflict detection task rather than general-purpose data cleaning: the goal is to identify records for which the identifiers match but the associated labels are inconsistent, without revealing the remaining private records. Existing PCDC protocols are mainly designed for direct two-party settings where data owners must remain online and participate in the main secure computation. To reduce this online burden, we propose SS-PCDC, a secret sharing-based PCDC framework in a cloud-assisted setting. Clients locally preprocess and secret-share their labeled datasets with two non-colluding cloud servers, which perform element matching, label consistency checking, and conflict detection over secret shares. Hash-based binning is used to reduce unnecessary secure comparisons. We instantiate the framework with two concrete protocols based on arithmetic secret sharing and Boolean secret sharing, respectively. We further extend exact PCDC to threshold-based fuzzy label conflict detection and propose SS-FPCDC, which reports a matched record as conflicting when the Hamming distance between its labels exceeds a public threshold. Security analyses show that the proposed protocols securely realize their corresponding ideal functionalities against a static semi-honest adversary corrupting at most one cloud server. Experimental results demonstrate the efficiency and scalability of SS-PCDC in its intended cloud-assisted setting, particularly for large datasets and longer labels.

With the increasing availability of large-scale spatiotemporal data from location-based services, trajectory publishing has become essential for data-driven analysis and intelligent applications. However, insufficient protection of trajectory location data may result in the disclosure of user privacy and social relationship information. To address this issue, we propose a semantic-aware privacy-preserving trajectory data publishing scheme (SAPP). First, a sliding-window algorithm is employed to extract stay points as key semantic locations and to generate a uniformly sampled set of candidate obfuscation points. Then, a semantic-aware scoring function is designed to probabilistically select candidate points that preserve semantics while avoiding sensitive regions. Furthermore, SAPP computes the sensitivity of each location based on semantic frequency and dynamically allocates the privacy budget. Finally, random noise is added to candidate trajectories using the Laplace mechanism. Through a dual-perturbation mechanism, spatial correlations in sensitive regions are weakened. Security analysis and experimental results further demonstrate that, compared with existing approaches, SAPP reduces TPPS and SFRR by up to 18% and 14%, respectively, indicating stronger resistance against trajectory inference and semantic leakage attacks while maintaining high data utility and time efficiency.

To address security risks such as single point of failure caused by traditional trusted authority, this paper proposes a distributed anonymous authentication scheme. This scheme not only ensures anonymous authentication for vehicle users to protect their privacy but also supports full traceability. Authorized vehicles can access corresponding computation services once authenticated. Specifically, to prevent single point of failure, a distributed architecture is adopted where two entities replace a single center, ensuring that neither party can perform anonymous authentication or tracing alone. Through secure collaboration between the servers, along with a commitment mechanism and elliptic curve cryptography, the anonymity of vehicle users is reliably maintained. Secondly, the scheme has a dual verification mechanism: the vehicle first passes cloud authentication, and the verified vehicle can enjoy the secure computing services provided by two clouds. If necessary, legitimate vehicles can also undergo identity verification to ensure normal communication between them. Through theoretical analysis and experimental verification, the security and feasibility of the scheme have been demonstrated. The quick reconnection mechanism we established reduces re authentication time costs by roughly 53 percent, while the dual server architecture supports privacy computing for vehicles after authentication.

The effective integration of multi-sourced data is crucial for enhancing decision-making capabilities in complex systems such as intelligent healthcare and finance. The XGBoost model, known for its high precision, is an ideal model for such fused data. However, privacy regulations and risks restrict the centralized use of multi-party data. To overcome this issue, we propose MKXGB, a secure, non-interactive XGBoost training scheme based on multi-key homomorphic encryption (MKHE), which can safely integrate encrypted multi-sourced data with the model training process. MKXGB allows multiple users to encrypt their data locally with their private keys. Subsequently, the Cloud Server (CS) can complete the model training directly on the aggregated ciphertexts without decryption or frequent interactions. We have designed secure algorithms for core operations such as gradient and gain computation. A key advantage of MKXGB is its unified support for both horizontal and vertical data partitioning within a single framework, allowing it to adapt to diverse multi-source data scenarios. Furthermore, users can go offline after a single round of data upload, supporting dynamic participation. Experiments show that MKXGB maintains accuracy comparable to plaintext training while ensuring strong privacy protection and reducing communication overhead, providing possibilities for future applications in privacy-sensitive scenarios.

Cross-modal retrieval is a technique that uses one modality to query another modality in multimedia data (e.g., retrieving images based on text, or retrieving text based on images). It can break down the barriers between different modalities and achieve seamless information connection. Secure cross-modal retrieval (SCMR) focuses on privacy issues in cross-modal retrieval, including private data of data owners (DOs) and private query requests of users. Current work on SCMR protects private information through homomorphic encryption, which makes the efficiency of the retrieval phase not ideal. Therefore, the conflict between retrieval efficiency and security has become an important issue that needs to be resolved in SCMR. We propose a scheme to achieve SCMR in the form of secret sharing in the IoT environment. In the scheme, the DO can secretly divide all the original data into two parts and upload them to two noncollusive cloud servers, respectively. The servers store the data and provide cross-modal retrieval for users. The security of the scheme is proved under semi-honest model, and the experiments show that our scheme is more efficient than previous work in the search phase. When the query dimension is 512 and the number of latent factors is 500, the search time is reduced by more than half compared with previous work.

With the rapid proliferation of vehicular technology, location-based services (LBS) have become a crucial component of Internet of Vehicles (IoV) applications, such as map navigation and health tracking. These applications rely on users' location information to provide services, enabling users to effectively share their locations, access information about nearbyactivities, and engage in real-time communication. However, the extensive collection and sharing of location data pose serious challenges to the semantic privacy preservation of user locations. To address these challenges in IoV, we propose a Semantic Correlation Trajectory Privacy-Preserving mechanism (SCTP). The SCTP combines the Hidden Markov Models (HMM) with differential privacy, aiming to protect the semantic privacy of user trajectory locations while maintaining high-quality location services and data usability. Our scheme introduces a trajectory prediction algorithm based on HMM, which dynamically and accurately predicts user trajectories and generates highly available semantically correlated trajectory datasets. Additionally, we design a personalized privacy budget allocation strategy based on semantic frequency. By assigning privacy weights, we significantly improve the usability of trajectory data while protecting data privacy. Theoretical analysis and experimental validation demonstrate that SCTP rigorously adheres to $\varepsilon$ -differential privacy standards while exhibiting significant advantages in safeguarding the semantic privacy of user locations.

In the e-Healthcare ecosystem, medical institutions increasingly rely on cloud computing platforms to outsource data storage and query processing, aiming to optimize service delivery efficiency. A critical component of such services is the top-$k$ query, which identifies the $k$ most relevant or highest-ranked records within datasets. However, since medical data contain sensitive patient information and require privacy-preserving outsourcing, traditional top-$k$ query schemes are no longer suitable, while existing privacy-preserving solutions suffer from high computational overhead in practice. To address these issues, we introduce a lightweight privacy-preserving secure top-$k$ query scheme. Specifically, our proposed scheme utilizes lightweight cryptographic tools, additive secret sharing and Function Secret Sharing (FSS) techniques, as the foundation of the underlying distributed secure computation. These techniques significantly reduce computational overhead while ensuring the privacy of medical data. Furthermore, we propose a Secure Max Heap Sorting (MH) protocol, which helps us to rapidly implement the top-$k$ query functionality in our scheme. Additionally, we design a set of fundamental secure protocols based on FSS, including the Secure Minimum Value (MinV) protocol, Secure Maximum Value (MaxV) protocol and Secure Heap Adjustment (HA) protocol. By integrating our cryptographic protocols with a secure squared Euclidean distance protocol, we construct a secure top-$k$ query scheme for e-Healthcare scenarios. Finally, we present formal security proofs under the semi-honest adversary model, which theoretically establish the security of the proposed scheme. Thanks to the adoption of secret sharing techniques, our scheme requires the client to only split the data into secret shares, a process that incurs nearly zero computational cost. The superior efficiency of our solution is further demonstrated through theoretical analysis and experimental evaluations.

The convergence of machine learning and searchable encryption enhances the ability to protect the privacy and security of data and enhances the processing power of confidential data. To enable users to efficiently perform machine learning tasks on encrypted data domains, we delve into oblivious keyword search with authorization (OKSA). The OKSA scheme effectively maintains the privacy of the user’s query keywords and prevents the cloud server from inferring ciphertext information through the searching process. However, limitations arise because the traditional OKSA approach does not support multi-keyword searches. If a data file is associated with multiple keywords, each keyword and corresponding data must be encrypted one by one, resulting in inefficiency. We introduce an innovative approach aimed at enhancing the efficiency of search processes while addressing the limitation of current encryption and search systems that handle only a single keyword. This method, known as the oblivious multiple keyword search with authorization (OMKSA), is designed for more effective keyword retrieval. One of our important innovations is that it uses the arithmetic techniques of bilinear pairs to generate new tokens and new search methods to optimize communication efficiency. Moreover, we present a detailed and rigorous demonstration of the security for our proposed protocol, aligned with the predefined security model. We conducted a comparative experiment to determine which of the two schemes, OKSA and OMKSA, is more efficient when querying multiple keywords. Based on our experimental results, our OMKSA is very efficient for data searchers. As the number of query keywords increases, the computational overhead of connected keyword searches remains stable. Finally, as we move into the 5G era, the potential applications of OMKSA are huge, with clear implications for areas such as machine learning and artificial intelligence. Our findings pave the way for further exploration and deployment of these frontier areas.

Data integrity auditing with data deduplication allows the cloud to store only one copy of the identical file while ensuring the integrity of outsourced data. To facilitate flexible updates of outsourced data, data integrity auditing schemes supporting data dynamics and deduplication have been proposed. However, existing schemes either impose significant computation and communication burden to achieve data dynamics while ensuring data integrity and deduplication, or incur substantial computation overhead during the phases of authenticator generation and auditing. To address the above problems, in this paper, we construct a secure deduplication and efficient data integrity auditing scheme with data dynamics for cloud storage (DIADD). We design a lightweight authenticator structure to produce data authenticators for data integrity auditing, which can achieve authenticator deduplication and greatly reduce the computation overhead in the authenticator generation phase. Additionally, the time-consuming operations can be eliminated in the auditing phase. To enhance the efficiency of data dynamics, we employ the multi-set hash function technology to produce the file tags. This allows data owners to compute a new file tag without needing to recover the entire original file when performing dynamic operations. Furthermore, security analysis and experimental results demonstrate that DIADD is both secure and efficient.