Secure and Verifiable Outsourcing of Large-Scale Nonnegative Matrix Factorization (NMF)

Duan, Jia; Zhou, Jiantao; Li, Yuanman

doi:10.1109/tsc.2019.2911282

Cited by 12 publications

(5 citation statements)

References 52 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To address these problems, a large number of improved solutions were proposed. For example, the VC protocols for securely outsourcing linear programming [24,25], matrix inversion [6,11], matrix-matrix multiplication [7,26], matrix determinant [8], linear regression [27,28], the large-scale system of linear equations [10,[29][30][31][32][33][34][35], compressed sensing reconstruction [12], and non-negative matrix factorization [36] have been investigated. Tang et al [37] presented a methodology called PILE for privacy-preserving federated learning with verifiable perturbations.…”

Section: Related Workmentioning

confidence: 99%

Enhancing Privacy Preservation in Verifiable Computation through Random Permutation Masking to Prevent Leakage

Yang,

Song

2023

Information

View full text Add to dashboard Cite

Outsourcing computation has become increasingly popular due to its cost-effectiveness, enabling users with limited resources to conduct large-scale computations on potentially untrusted cloud platforms. In order to safeguard privacy, verifiable computing (VC) has emerged as a secure approach, ensuring that the cloud cannot discern users’ input and output. Random permutation masking (RPM) is a widely adopted technique in VC protocols to provide robust privacy protection. This work presents a precise definition of the privacy-preserving property of RPM by employing indistinguishability experiments. Moreover, an innovative attack exploiting the greatest common divisor and the least common multiple of each row and column in the encrypted matrices is introduced against RPM. Unlike previous density-based attacks, this novel approach offers a significant advantage by allowing the reconstruction of matrix values from the ciphertext based on RPM. A comprehensive demonstration was provided to illustrate the failure of protocols based on RPM in maintaining the privacy-preserving property under this proposed attack. Furthermore, an extensive series of experiments is conducted to thoroughly validate the effectiveness and advantages of the attack against RPM. The findings of this research highlight vulnerabilities in RPM-based VC protocols and underline the pressing need for further enhancements and alternative privacy-preserving mechanisms in outsourcing computation.

show abstract

Section: Related Workmentioning

confidence: 99%

Enhancing Privacy Preservation in Verifiable Computation through Random Permutation Masking to Prevent Leakage

Yang,

Song

2023

Information

View full text Add to dashboard Cite

show abstract

“…In addition, Nonnegative Matrix Factorization (NMF) is widely used in DIP, face recognition, text analysis, and other fields. Thus, there are many secure outsourcing algorithms for NMF [27][28][29][30]. Matrix inverse is also one of the most basic computations in large-scale data analysis.…”

Section: Related Work and Comparative Analysismentioning

confidence: 99%

Secure Outsourcing of Matrix Determinant Computation under the Malicious Cloud

Song

Sang

2021

Sensors

View full text Add to dashboard Cite

Computing the determinant of large matrix is a time-consuming task, which is appearing more and more widely in science and engineering problems in the era of big data. Fortunately, cloud computing can provide large storage and computation resources, and thus, act as an ideal platform to complete computation outsourced from resource-constrained devices. However, cloud computing also causes security issues. For example, the curious cloud may spy on user privacy through outsourced data. The malicious cloud violating computing scripts, as well as cloud hardware failure, will lead to incorrect results. Therefore, we propose a secure outsourcing algorithm to compute the determinant of large matrix under the malicious cloud mode in this paper. The algorithm protects the privacy of the original matrix by applying row/column permutation and other transformations to the matrix. To resist malicious cheating on the computation tasks, a new verification method is utilized in our algorithm. Unlike previous algorithms that require multiple rounds of verification, our verification requires only one round without trading off the cheating detectability, which greatly reduces the local computation burden. Both theoretical and experimental analysis demonstrate that our algorithm achieves a better efficiency on local users than previous ones on various dimensions of matrices, without sacrificing the security requirements in terms of privacy protection and cheating detectability.

show abstract

“…Zhang et al [31] proposed a secure outsourcing computation framework for PCA-based face recognition. Duan et al [8] and Fu et al [10] proposed outsourcing computation frameworks for non-negative matrix factorization through random permutation and homomorphic encryption, respectively. Luo et al [22] proposed a masking based outsourcing computation method for QR and LU factorization.…”

Section: Related Workmentioning

confidence: 99%

Practical Lossless Federated Singular Vector Decomposition over Billion-Scale Data

Chai¹,

Wang²,

Zhang³

et al. 2021

Preprint

View full text Add to dashboard Cite

With the promulgation of data protection laws (e.g., GDPR in 2018), privacy preservation has become a general agreement in applications where cross-domain sensitive data are utilized. Out of many privacy-preserving techniques, federated learning (FL) has received much attention as a bridge for secure data connection and cooperation. Although FL's research works have surged, some classical data modeling methods are not well accommodated in FL. In this paper, we propose the first masking-based federated singular vector decomposition method, called FedSVD. FedSVD protects the raw data through a singular value invariance mask, which can be further removed from the SVD results. Compared with prior privacy-preserving SVD solutions, FedSVD has lossless results, high confidentiality, and excellent scalability. We provide privacy proof showing that FedSVD has guaranteed data confidentiality. Empirical experiments on real-life datasets and synthetic data have verified the effectiveness of our method. The reconstruction error of FedSVD is around 0.000001% of the raw data, validating the lossless property of FedSVD. The scalability of FedSVD is nearly the same as the standalone SVD algorithm. Hence, FedSVD can bring privacy protection almost without sacrificing any computation time or communication overhead.

show abstract

Secure and Verifiable Outsourcing of Large-Scale Nonnegative Matrix Factorization (NMF)

Cited by 12 publications

References 52 publications

Enhancing Privacy Preservation in Verifiable Computation through Random Permutation Masking to Prevent Leakage

Enhancing Privacy Preservation in Verifiable Computation through Random Permutation Masking to Prevent Leakage

Secure Outsourcing of Matrix Determinant Computation under the Malicious Cloud

Practical Lossless Federated Singular Vector Decomposition over Billion-Scale Data

Contact Info

Product

Resources

About