On Nov.4^(th), AQSIQ (General Administration of Quality Supervision,Inspection and Quarantine of the People' s Republic of China), SAC (Standardization Administrationof China), National Audit Office of China (CNAO...On Nov.4^(th), AQSIQ (General Administration of Quality Supervision,Inspection and Quarantine of the People' s Republic of China), SAC (Standardization Administrationof China), National Audit Office of China (CNAO), and National Ministry of Finance of China jointlyheld the conference press on the national standard of Information Technology--Data Interface ofAccounting Software (GB/T 19581-2004) in Beijing. The standard was approved and issued on Sept. 20,2004 by AQSIQ and SAC, and it would come into effect all over the whole nation from January 1^(st),2005. Pu Changcheng, Vice Director of AQSIQ, Shi Aizhong, Vice Director of CNAO, Li Zhonghai. amember of the Party Group of AQSIQ and Director of SAC, the other leaders of concerned departmentssuch as National Ministry of Finance, National Telegraphy Office, and etc. attended the ConferencePress and made speeches. They fully affirmed the important significance and the achievements onstandardization work of electronic government business, and also they set new demands on the workfor the future.展开更多
With the widespread data collection and processing,privacy-preserving machine learning has become increasingly important in addressing privacy risks related to individuals.Support vector machine(SVM)is one of the most...With the widespread data collection and processing,privacy-preserving machine learning has become increasingly important in addressing privacy risks related to individuals.Support vector machine(SVM)is one of the most elementary learning models of machine learning.Privacy issues surrounding SVM classifier training have attracted increasing attention.In this paper,we investigate Differential Privacy-compliant Federated Machine Learning with Dimensionality Reduction,called FedDPDR-DPML,which greatly improves data utility while providing strong privacy guarantees.Considering in distributed learning scenarios,multiple participants usually hold unbalanced or small amounts of data.Therefore,FedDPDR-DPML enables multiple participants to collaboratively learn a global model based on weighted model averaging and knowledge aggregation and then the server distributes the global model to each participant to improve local data utility.Aiming at high-dimensional data,we adopt differential privacy in both the principal component analysis(PCA)-based dimensionality reduction phase and SVM classifiers training phase,which improves model accuracy while achieving strict differential privacy protection.Besides,we train Differential privacy(DP)-compliant SVM classifiers by adding noise to the objective function itself,thus leading to better data utility.Extensive experiments on three high-dimensional datasets demonstrate that FedDPDR-DPML can achieve high accuracy while ensuring strong privacy protection.展开更多
Multidimensional data provides enormous opportunities in a variety of applications. Recent research has indicated the failure of existing sanitization techniques (e.g., k-anonymity) to provide rigorous privacy guara...Multidimensional data provides enormous opportunities in a variety of applications. Recent research has indicated the failure of existing sanitization techniques (e.g., k-anonymity) to provide rigorous privacy guarantees. Privacy- preserving multidimensional data publishing currently lacks a solid theoretical foundation. It is urgent to develop new techniques with provable privacy guarantees, e-Differential privacy is the only method that can provide such guarantees. In this paper, we propose a multidimensional data publishing scheme that ensures c-differential privacy while providing accurate results for query processing. The proposed solution applies nonstandard wavelet transforms on the raw multidimensional data and adds noise to guarantee c-differential privacy. Then, the scheme processes arbitrarily queries directly in the noisy wavelet- coefficient synopses of relational tables and expands the noisy wavelet coefficients back into noisy relational tuples until the end result of the query. Moreover, experimental results demonstrate the high accuracy and effectiveness of our approach.展开更多
Speech data publishing breaches users'data privacy,thereby causing more privacy disclosure.Existing work sanitizes content,voice,and voiceprint of speech data without considering the consistence among these three ...Speech data publishing breaches users'data privacy,thereby causing more privacy disclosure.Existing work sanitizes content,voice,and voiceprint of speech data without considering the consistence among these three features,and thus is susceptible to inference attacks.To address the problem,we design a privacy-preserving protocol for speech data publishing(P3S2)that takes the corrections among the three factors into consideration.To concrete,we first propose a three-dimensional sanitization that uses feature learning to capture characteristics in each dimension,and then sanitize speech data using the learned features.As a result,the correlations among the three dimensions of the sanitized speech data are guaranteed.Furthermore,the(ε,δ)-differential privacy is used to theoretically prove both the data privacy preservation and the data utility guarantee of P3S2,filling the gap of algorithm design and performance evaluation.Finally,simulations on two real world datasets have demonstrated both the data privacy preservation and the data utility guarantee.展开更多
Data is not only a key production factor but also an important foundation and strategic resource that drives economic growth and social progress in the era of digital economy. Data sharing and innovative utilization i...Data is not only a key production factor but also an important foundation and strategic resource that drives economic growth and social progress in the era of digital economy. Data sharing and innovative utilization in an ethical and responsible manner is a focus of the current studies on smart city construction. Taking Shenzhen as an example, this paper analyzes the three typical cases of data legislation, data sharing and utilization,and data-based anti-epidemic action in its smart city construction and explores the respective role of the four actors of the government, enterprises,research institutes, and the public in innovating data utilization to serve the public interests through data sharing. By studying Shenzhen’s multi-actor interaction mechanism of smart city construction, the paper tries to provide a useful experience for the construction of smart cities in China from the perspectives of data management, data sharing, and innovative data utilization.展开更多
This paper addresses a special and imperceptible class of privacy,called implicit privacy.In contrast to traditional(explicit)privacy,implicit privacy has two essential prop-erties:(1)It is not initially defined as a ...This paper addresses a special and imperceptible class of privacy,called implicit privacy.In contrast to traditional(explicit)privacy,implicit privacy has two essential prop-erties:(1)It is not initially defined as a privacy attribute;(2)it is strongly associated with privacy attributes.In other words,attackers could utilize it to infer privacy attributes with a certain probability,indirectly resulting in the disclosure of private information.To deal with the implicit privacy disclosure problem,we give a measurable definition of implicit privacy,and propose an ex-ante implicit privacy-preserving framework based on data generation,called IMPOSTER.The framework consists of an implicit privacy detection module and an implicit privacy protection module.The former uses normalized mutual information to detect implicit privacy attributes that are strongly related to traditional privacy attributes.Based on the idea of data generation,the latter equips the Generative Adversarial Network(GAN)framework with an additional discriminator,which is used to eliminate the association between traditional privacy attributes and implicit ones.We elaborate a theoretical analysis for the convergence of the framework.Experiments demonstrate that with the learned gen-erator,IMPOSTER can alleviate the disclosure of implicit privacy while maintaining good data utility.展开更多
This paper addresses a special and imperceptible class of privacy,called implicit privacy.In contrast to traditional(explicit)privacy,implicit privacy has two essential properties:(1)It is not initially de ned as a pr...This paper addresses a special and imperceptible class of privacy,called implicit privacy.In contrast to traditional(explicit)privacy,implicit privacy has two essential properties:(1)It is not initially de ned as a privacy attribute;(2)it is strongly associated with privacy attributes.In other words,attackers could utilize it to infer privacy attributes with a certain probability,indirectly resulting in the disclosure of private information.To deal with the implicit privacy disclosure problem,we give a measurable de nition of implicit privacy,and propose an ex-ante implicit privacy-preserving framework based on data generation,called IMPOSTER.The framework consists of an implicit privacy detection module and an implicit privacy protection module.The former uses normalized mutual information to detect implicit privacy attributes that are strongly related to traditional privacy attributes.Based on the idea of data generation,the latter equips the Generative Adversarial Network(GAN)framework with an additional discriminator,which is used to eliminate the association between traditional privacy attributes and implicit ones.We elaborate a theoretical analysis for the convergence of the framework.Experiments demonstrate that with the learned generator,IMPOSTER can alleviate the disclosure of implicit privacy while maintaining good data utility.展开更多
文摘On Nov.4^(th), AQSIQ (General Administration of Quality Supervision,Inspection and Quarantine of the People' s Republic of China), SAC (Standardization Administrationof China), National Audit Office of China (CNAO), and National Ministry of Finance of China jointlyheld the conference press on the national standard of Information Technology--Data Interface ofAccounting Software (GB/T 19581-2004) in Beijing. The standard was approved and issued on Sept. 20,2004 by AQSIQ and SAC, and it would come into effect all over the whole nation from January 1^(st),2005. Pu Changcheng, Vice Director of AQSIQ, Shi Aizhong, Vice Director of CNAO, Li Zhonghai. amember of the Party Group of AQSIQ and Director of SAC, the other leaders of concerned departmentssuch as National Ministry of Finance, National Telegraphy Office, and etc. attended the ConferencePress and made speeches. They fully affirmed the important significance and the achievements onstandardization work of electronic government business, and also they set new demands on the workfor the future.
基金supported in part by National Natural Science Foundation of China(Nos.62102311,62202377,62272385)in part by Natural Science Basic Research Program of Shaanxi(Nos.2022JQ-600,2022JM-353,2023-JC-QN-0327)+2 种基金in part by Shaanxi Distinguished Youth Project(No.2022JC-47)in part by Scientific Research Program Funded by Shaanxi Provincial Education Department(No.22JK0560)in part by Distinguished Youth Talents of Shaanxi Universities,and in part by Youth Innovation Team of Shaanxi Universities.
文摘With the widespread data collection and processing,privacy-preserving machine learning has become increasingly important in addressing privacy risks related to individuals.Support vector machine(SVM)is one of the most elementary learning models of machine learning.Privacy issues surrounding SVM classifier training have attracted increasing attention.In this paper,we investigate Differential Privacy-compliant Federated Machine Learning with Dimensionality Reduction,called FedDPDR-DPML,which greatly improves data utility while providing strong privacy guarantees.Considering in distributed learning scenarios,multiple participants usually hold unbalanced or small amounts of data.Therefore,FedDPDR-DPML enables multiple participants to collaboratively learn a global model based on weighted model averaging and knowledge aggregation and then the server distributes the global model to each participant to improve local data utility.Aiming at high-dimensional data,we adopt differential privacy in both the principal component analysis(PCA)-based dimensionality reduction phase and SVM classifiers training phase,which improves model accuracy while achieving strict differential privacy protection.Besides,we train Differential privacy(DP)-compliant SVM classifiers by adding noise to the objective function itself,thus leading to better data utility.Extensive experiments on three high-dimensional datasets demonstrate that FedDPDR-DPML can achieve high accuracy while ensuring strong privacy protection.
基金the National Basic Research Program of China under Grant 2013CB338004,Doctoral Program of Higher Education of China under Grant No.20120073120034,National Natural Science Foundation of China under Grants No.61070204,61101108,and National S&T Major Program under Grant No.2011ZX03002-005-01
文摘Multidimensional data provides enormous opportunities in a variety of applications. Recent research has indicated the failure of existing sanitization techniques (e.g., k-anonymity) to provide rigorous privacy guarantees. Privacy- preserving multidimensional data publishing currently lacks a solid theoretical foundation. It is urgent to develop new techniques with provable privacy guarantees, e-Differential privacy is the only method that can provide such guarantees. In this paper, we propose a multidimensional data publishing scheme that ensures c-differential privacy while providing accurate results for query processing. The proposed solution applies nonstandard wavelet transforms on the raw multidimensional data and adds noise to guarantee c-differential privacy. Then, the scheme processes arbitrarily queries directly in the noisy wavelet- coefficient synopses of relational tables and expands the noisy wavelet coefficients back into noisy relational tuples until the end result of the query. Moreover, experimental results demonstrate the high accuracy and effectiveness of our approach.
基金National Natural Science Foundation of China(No.61902060)Shanghai Sailing Program,China(No.19YF1402100)Fundamental Research Funds for the Central Universities,China(No.2232019D3-51)。
文摘Speech data publishing breaches users'data privacy,thereby causing more privacy disclosure.Existing work sanitizes content,voice,and voiceprint of speech data without considering the consistence among these three features,and thus is susceptible to inference attacks.To address the problem,we design a privacy-preserving protocol for speech data publishing(P3S2)that takes the corrections among the three factors into consideration.To concrete,we first propose a three-dimensional sanitization that uses feature learning to capture characteristics in each dimension,and then sanitize speech data using the learned features.As a result,the correlations among the three dimensions of the sanitized speech data are guaranteed.Furthermore,the(ε,δ)-differential privacy is used to theoretically prove both the data privacy preservation and the data utility guarantee of P3S2,filling the gap of algorithm design and performance evaluation.Finally,simulations on two real world datasets have demonstrated both the data privacy preservation and the data utility guarantee.
基金supported by the National Natural Science Foundation of China(Project No.52078197)。
文摘Data is not only a key production factor but also an important foundation and strategic resource that drives economic growth and social progress in the era of digital economy. Data sharing and innovative utilization in an ethical and responsible manner is a focus of the current studies on smart city construction. Taking Shenzhen as an example, this paper analyzes the three typical cases of data legislation, data sharing and utilization,and data-based anti-epidemic action in its smart city construction and explores the respective role of the four actors of the government, enterprises,research institutes, and the public in innovating data utilization to serve the public interests through data sharing. By studying Shenzhen’s multi-actor interaction mechanism of smart city construction, the paper tries to provide a useful experience for the construction of smart cities in China from the perspectives of data management, data sharing, and innovative data utilization.
基金supported in part by the National Key Research and Development Program of China under Grant 2018YFB2100801in part by the National Natural Science Foundation of China(NSFC)under Grant 61972287in part by the Fundamental Research Funds for the Central Universities under Grant 22120210524.
文摘This paper addresses a special and imperceptible class of privacy,called implicit privacy.In contrast to traditional(explicit)privacy,implicit privacy has two essential prop-erties:(1)It is not initially defined as a privacy attribute;(2)it is strongly associated with privacy attributes.In other words,attackers could utilize it to infer privacy attributes with a certain probability,indirectly resulting in the disclosure of private information.To deal with the implicit privacy disclosure problem,we give a measurable definition of implicit privacy,and propose an ex-ante implicit privacy-preserving framework based on data generation,called IMPOSTER.The framework consists of an implicit privacy detection module and an implicit privacy protection module.The former uses normalized mutual information to detect implicit privacy attributes that are strongly related to traditional privacy attributes.Based on the idea of data generation,the latter equips the Generative Adversarial Network(GAN)framework with an additional discriminator,which is used to eliminate the association between traditional privacy attributes and implicit ones.We elaborate a theoretical analysis for the convergence of the framework.Experiments demonstrate that with the learned gen-erator,IMPOSTER can alleviate the disclosure of implicit privacy while maintaining good data utility.
基金the National Key Research and Development Program of China under Grant 2018YFB2100801in part by the National Natural Science Foundation of China(NSFC)under Grant 61972287in part by the Fundamental Research Funds for the Central Universities under Grant 22120210524.
文摘This paper addresses a special and imperceptible class of privacy,called implicit privacy.In contrast to traditional(explicit)privacy,implicit privacy has two essential properties:(1)It is not initially de ned as a privacy attribute;(2)it is strongly associated with privacy attributes.In other words,attackers could utilize it to infer privacy attributes with a certain probability,indirectly resulting in the disclosure of private information.To deal with the implicit privacy disclosure problem,we give a measurable de nition of implicit privacy,and propose an ex-ante implicit privacy-preserving framework based on data generation,called IMPOSTER.The framework consists of an implicit privacy detection module and an implicit privacy protection module.The former uses normalized mutual information to detect implicit privacy attributes that are strongly related to traditional privacy attributes.Based on the idea of data generation,the latter equips the Generative Adversarial Network(GAN)framework with an additional discriminator,which is used to eliminate the association between traditional privacy attributes and implicit ones.We elaborate a theoretical analysis for the convergence of the framework.Experiments demonstrate that with the learned generator,IMPOSTER can alleviate the disclosure of implicit privacy while maintaining good data utility.