An efficient approach for secure multi-party computation without authenticated channel

Vu, Duy-Hien; Luong, The-Dung; Ho, Tu Bao

doi:10.1016/j.ins.2019.07.031

Cited by 22 publications

(16 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Access control functions (e.g., [11]) achieves privacy protection through authorization models and access control policy operations. Encryption functions applies encryption mechanisms on data values (e.g., Secure Multi-party Computation [22]).…”

Section: Privacy Protection Functionsmentioning

confidence: 99%

P-SGD: A Stochastic Gradient Descent Solution for Privacy-Preserving During Protection Transitions

Bou-Chaaya

Chbeir

Barhamgi

et al. 2021

Advanced Information Systems Engineering

View full text Add to dashboard Cite

Advances in privacy-enhancing technologies, such as contextaware and personalized privacy models, have paved the way for successful management of the data utility-privacy trade-off. However, significantly lowering the level of data protection when balancing utility-privacy to meet the individual's needs makes subsequent protected data more precise. This increases the adversary's ability to reveal the real values of the previous correlated data that needed more protection, making existing privacy models vulnerable to inference attacks. To overcome this problem, we propose in this paper a stochastic gradient descent solution for privacy-preserving during protection transitions, denoted P-SGD. The goal of this solution is to minimize the precision gap between sequential data when downshifting the protection by the privacy model. P-SGD intervenes at the protection descent phase and performs an iterative process that measures data dependencies, and gradually reduces protection accordingly until the desired protection level is reached. It considers also possible changes in protection functions and studies their impact on the protection descent rate. We validated our proposal and evaluated its performance. The results show that P-SGD is fast, scalable, and maintains low computational and storage complexity.

show abstract

Section: Privacy Protection Functionsmentioning

confidence: 99%

P-SGD: A Stochastic Gradient Descent Solution for Privacy-Preserving During Protection Transitions

Bou-Chaaya

Chbeir

Barhamgi

et al. 2021

Advanced Information Systems Engineering

View full text Add to dashboard Cite

show abstract

“…We accomplished substantial experiments to signify the effectiveness of our proposed methodologies by comparing them with subsisting technologies. For that purpose, we selected data set of wine quality [19]. This data set is often used in many research papers which are related to multiparty computation for protecting the privacy in distributed information mining task.…”

Section: Performance Comparabilitymentioning

confidence: 99%

“…This data set is often used in many research papers which are related to multiparty computation for protecting the privacy in distributed information mining task. Two experiments were accomplished through our proposed methodology for comparison with the outcome of experiments in [19,20]. Outcomes of experiments are shown in Table 6.…”

Section: Performance Comparabilitymentioning

confidence: 99%

An Intelligent Approach for Protecting Privacy in Distributed Information Mining Using Secured Computation of Multiple Participating Sites

Sharma¹,

Bhushan²,

Boahar³

et al. 2021

ISI

View full text Add to dashboard Cite

Information mining is a very well-known task using which managers can take a better decision regarding data operations. For that purpose, they need to get useful information from a large amount of raw data. This kind of big data mining is usually carried out on unstructured data that is huge in context of its size. Due to the vast size, the data mining process faces confidentiality and security breaches issues. There are many technologies through which we can search for useful information for getting fruitful results towards the fulfilment of organizational goals. However, it's a big challenge for researchers to get knowledge from the large amount of data that is owned by multiple parties which are located at different sites. For that, they need to perform Distributed Data Mining task with the concern of privacy leakage. While secure multiparty computation presents the solution to this problem but there are some issues which are untouched yet. In this paper, we presented two data privacy issues not even solved by multiparty computation. The presented algorithms are designed using the process of matrix computation with the help of encoding and decomposition methods. The implementation of the work is based on a secure Multi-Party Computation (MPC) protocol named SPDZ so that we can perform the sharing of data values. We analyzed the results produced by single machine and proposed design and implementation to show the similarity between both. Experimental results show the effectiveness of implemented algorithms and their implementation to preserve the privacy of distributed data in the process of distributed data mining.

show abstract

“…In this way, a (computationally) secure sum can be performed without the need for pairwise secure channels. Other more recently published protocols [18,27], fulfilling the property of collusion-resistance for more than two parties, build on homomorphic encryption techniques (HE). HE allows additions or multiplications of ciphertexts on any two numbers m 1 and m 2 that have been encrypted by an encryption function Enc(x).…”

Section: Secure Sum Protocolsmentioning

confidence: 99%

“…We think that this is feasible in practice. A trend of secure sum research is to guarantee security without secure channels thanks to HE [14,27]. But as it is easy to implement secure paired connections using established TLS connections, the need of being able to communicate without secure channels in complex computations as collaborative DL might be rare.…”

Section: Introductionmentioning

confidence: 99%

Secure Sum Outperforms Homomorphic Encryption in (Current) Collaborative Deep Learning

Boer¹,

Krämer²

2020

Preprint

View full text Add to dashboard Cite

Deep learning (DL) approaches are achieving extraordinary results in a wide range of domains but often require a massive collection of private data. Hence, methods for training neural networks on the joint data of different data owners, that keep each party's input confidential, are called for. We address the setting of horizontally distributed data in deep learning, where the participants' vulnerable intermediate results have to be processed in a privacy-preserving manner. The predominant scheme for this setting is based on homomorphic encryption (HE), and it is widely considered to be without alternative. In contrast to this, we demonstrate that a carefully chosen, less complex and computationally less expensive secure sum protocol in conjunction with default secure channels exhibits superior properties in terms of both collusion-resistance and runtime. Finally, we discuss several open research questions in the context of collaborative DL, which possibly might lead back to HE-based solutions.

show abstract

An efficient approach for secure multi-party computation without authenticated channel

Cited by 22 publications

References 16 publications

P-SGD: A Stochastic Gradient Descent Solution for Privacy-Preserving During Protection Transitions

P-SGD: A Stochastic Gradient Descent Solution for Privacy-Preserving During Protection Transitions

An Intelligent Approach for Protecting Privacy in Distributed Information Mining Using Secured Computation of Multiple Participating Sites

Secure Sum Outperforms Homomorphic Encryption in (Current) Collaborative Deep Learning

Contact Info

Product

Resources

About