Compressive mechanism

Utilizing sparse representation in differential privacy

Yang D. Li, Zhenjie Zhang, Marianne Winslett, Yin Yang

Research output: Chapter in Book/Report/Conference proceedingConference contribution

31 Citations (Scopus)

Abstract

Differential privacy provides the first theoretical foundation with provable privacy guarantee against adversaries with arbitrary prior knowledge. The main idea to achieve differential privacy is to inject random noise into statistical query results. Besides correctness, the most important goal in the design of a differentially private mechanism is to reduce the effect of random noise, ensuring that the noisy results can still be useful. This paper proposes the compressive mechanism, a novel solution on the basis of state-of-the-art compression technique, called compressive sensing. Compressive sensing is a decent theoretical tool for compact synopsis construction, using random projections. In this paper, we show that the amount of noise is significantly reduced from O(√n) to O(log(n)), when the noise insertion procedure is carried on the synopsis samples instead of the original database. As an extension, we also apply the proposed compressive mechanism to solve the problem of continual release of statistical results. Extensive experiments using real datasets justify our accuracy claims.

Original languageEnglish
Title of host publicationWPES'11 - Proceedings of the 10th Annual ACM Workshop on Privacy in the Electronic Society
Pages177-182
Number of pages6
DOIs
Publication statusPublished - 2011
Externally publishedYes
Event10th Annual ACM Workshop on Privacy in the Electronic Society, WPES'11 - Co-located with 18th ACM Conference on Computer and Communications Security, CCS 2011 - Chicago, IL, United States
Duration: 17 Oct 201117 Oct 2011

Other

Other10th Annual ACM Workshop on Privacy in the Electronic Society, WPES'11 - Co-located with 18th ACM Conference on Computer and Communications Security, CCS 2011
CountryUnited States
CityChicago, IL
Period17/10/1117/10/11

Fingerprint

Experiments

Keywords

  • Compressive sensing
  • Differential privacy
  • Randomness

ASJC Scopus subject areas

  • Software
  • Computer Networks and Communications

Cite this

Li, Y. D., Zhang, Z., Winslett, M., & Yang, Y. (2011). Compressive mechanism: Utilizing sparse representation in differential privacy. In WPES'11 - Proceedings of the 10th Annual ACM Workshop on Privacy in the Electronic Society (pp. 177-182) https://doi.org/10.1145/2046556.2046581

Compressive mechanism : Utilizing sparse representation in differential privacy. / Li, Yang D.; Zhang, Zhenjie; Winslett, Marianne; Yang, Yin.

WPES'11 - Proceedings of the 10th Annual ACM Workshop on Privacy in the Electronic Society. 2011. p. 177-182.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Li, YD, Zhang, Z, Winslett, M & Yang, Y 2011, Compressive mechanism: Utilizing sparse representation in differential privacy. in WPES'11 - Proceedings of the 10th Annual ACM Workshop on Privacy in the Electronic Society. pp. 177-182, 10th Annual ACM Workshop on Privacy in the Electronic Society, WPES'11 - Co-located with 18th ACM Conference on Computer and Communications Security, CCS 2011, Chicago, IL, United States, 17/10/11. https://doi.org/10.1145/2046556.2046581
Li YD, Zhang Z, Winslett M, Yang Y. Compressive mechanism: Utilizing sparse representation in differential privacy. In WPES'11 - Proceedings of the 10th Annual ACM Workshop on Privacy in the Electronic Society. 2011. p. 177-182 https://doi.org/10.1145/2046556.2046581
Li, Yang D. ; Zhang, Zhenjie ; Winslett, Marianne ; Yang, Yin. / Compressive mechanism : Utilizing sparse representation in differential privacy. WPES'11 - Proceedings of the 10th Annual ACM Workshop on Privacy in the Electronic Society. 2011. pp. 177-182
@inproceedings{c9eb3dd17fd543a4be74e576c25b5e52,
title = "Compressive mechanism: Utilizing sparse representation in differential privacy",
abstract = "Differential privacy provides the first theoretical foundation with provable privacy guarantee against adversaries with arbitrary prior knowledge. The main idea to achieve differential privacy is to inject random noise into statistical query results. Besides correctness, the most important goal in the design of a differentially private mechanism is to reduce the effect of random noise, ensuring that the noisy results can still be useful. This paper proposes the compressive mechanism, a novel solution on the basis of state-of-the-art compression technique, called compressive sensing. Compressive sensing is a decent theoretical tool for compact synopsis construction, using random projections. In this paper, we show that the amount of noise is significantly reduced from O(√n) to O(log(n)), when the noise insertion procedure is carried on the synopsis samples instead of the original database. As an extension, we also apply the proposed compressive mechanism to solve the problem of continual release of statistical results. Extensive experiments using real datasets justify our accuracy claims.",
keywords = "Compressive sensing, Differential privacy, Randomness",
author = "Li, {Yang D.} and Zhenjie Zhang and Marianne Winslett and Yin Yang",
year = "2011",
doi = "10.1145/2046556.2046581",
language = "English",
isbn = "9781450310024",
pages = "177--182",
booktitle = "WPES'11 - Proceedings of the 10th Annual ACM Workshop on Privacy in the Electronic Society",

}

TY - GEN

T1 - Compressive mechanism

T2 - Utilizing sparse representation in differential privacy

AU - Li, Yang D.

AU - Zhang, Zhenjie

AU - Winslett, Marianne

AU - Yang, Yin

PY - 2011

Y1 - 2011

N2 - Differential privacy provides the first theoretical foundation with provable privacy guarantee against adversaries with arbitrary prior knowledge. The main idea to achieve differential privacy is to inject random noise into statistical query results. Besides correctness, the most important goal in the design of a differentially private mechanism is to reduce the effect of random noise, ensuring that the noisy results can still be useful. This paper proposes the compressive mechanism, a novel solution on the basis of state-of-the-art compression technique, called compressive sensing. Compressive sensing is a decent theoretical tool for compact synopsis construction, using random projections. In this paper, we show that the amount of noise is significantly reduced from O(√n) to O(log(n)), when the noise insertion procedure is carried on the synopsis samples instead of the original database. As an extension, we also apply the proposed compressive mechanism to solve the problem of continual release of statistical results. Extensive experiments using real datasets justify our accuracy claims.

AB - Differential privacy provides the first theoretical foundation with provable privacy guarantee against adversaries with arbitrary prior knowledge. The main idea to achieve differential privacy is to inject random noise into statistical query results. Besides correctness, the most important goal in the design of a differentially private mechanism is to reduce the effect of random noise, ensuring that the noisy results can still be useful. This paper proposes the compressive mechanism, a novel solution on the basis of state-of-the-art compression technique, called compressive sensing. Compressive sensing is a decent theoretical tool for compact synopsis construction, using random projections. In this paper, we show that the amount of noise is significantly reduced from O(√n) to O(log(n)), when the noise insertion procedure is carried on the synopsis samples instead of the original database. As an extension, we also apply the proposed compressive mechanism to solve the problem of continual release of statistical results. Extensive experiments using real datasets justify our accuracy claims.

KW - Compressive sensing

KW - Differential privacy

KW - Randomness

UR - http://www.scopus.com/inward/record.url?scp=80955157833&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=80955157833&partnerID=8YFLogxK

U2 - 10.1145/2046556.2046581

DO - 10.1145/2046556.2046581

M3 - Conference contribution

SN - 9781450310024

SP - 177

EP - 182

BT - WPES'11 - Proceedings of the 10th Annual ACM Workshop on Privacy in the Electronic Society

ER -