TY - GEN
T1 - MVG mechanism
T2 - 25th ACM Conference on Computer and Communications Security, CCS 2018
AU - Chanyaswad, Thee
AU - Dytso, Alex
AU - Poor, H. Vincent
AU - Mittal, Prateek
N1 - Funding Information:
The authors would like to thank the reviewers for their valuable feedback that helped improve the paper. This work was supported in part by the National Science Foundation (NSF) under Grants CNS-1553437, CCF-1617286, and CNS-1702808; an Army Research Office YIP Award; and faculty research awards from Google, Cisco, Intel, and IBM.
Publisher Copyright:
© 2018 Association for Computing Machinery.
PY - 2018/10/15
Y1 - 2018/10/15
N2 - Differential privacy mechanism design has traditionally been tailored for a scalar-valued query function. Although many mechanisms such as the Laplace and Gaussian mechanisms can be extended to a matrix-valued query function by adding i.i.d. noise to each element of the matrix, this method is often suboptimal as it forfeits an opportunity to exploit the structural characteristics typically associated with matrix analysis. To address this challenge, we propose a novel differential privacy mechanism called the Matrix-Variate Gaussian (MVG) mechanism, which adds a matrix-valued noise drawn from a matrix-variate Gaussian distribution, and we rigorously prove that the MVG mechanism preserves (?, d)-differential privacy. Furthermore, we introduce the concept of directional noise made possible by the design of the MVG mechanism. Directional noise allows the impact of the noise on the utility of the matrix-valued query function to be moderated. Finally, we experimentally demonstrate the performance of our mechanism using three matrix-valued queries on three privacy-sensitive datasets. We find that the MVG mechanism can notably outperforms four previous state-of-the-art approaches, and provides comparable utility to the non-private baseline.
AB - Differential privacy mechanism design has traditionally been tailored for a scalar-valued query function. Although many mechanisms such as the Laplace and Gaussian mechanisms can be extended to a matrix-valued query function by adding i.i.d. noise to each element of the matrix, this method is often suboptimal as it forfeits an opportunity to exploit the structural characteristics typically associated with matrix analysis. To address this challenge, we propose a novel differential privacy mechanism called the Matrix-Variate Gaussian (MVG) mechanism, which adds a matrix-valued noise drawn from a matrix-variate Gaussian distribution, and we rigorously prove that the MVG mechanism preserves (?, d)-differential privacy. Furthermore, we introduce the concept of directional noise made possible by the design of the MVG mechanism. Directional noise allows the impact of the noise on the utility of the matrix-valued query function to be moderated. Finally, we experimentally demonstrate the performance of our mechanism using three matrix-valued queries on three privacy-sensitive datasets. We find that the MVG mechanism can notably outperforms four previous state-of-the-art approaches, and provides comparable utility to the non-private baseline.
KW - Differential privacy
KW - Directional noise
KW - MVG mechanism
KW - Matrix-valued query
KW - Matrix-variate Gaussian
UR - http://www.scopus.com/inward/record.url?scp=85056896227&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85056896227&partnerID=8YFLogxK
U2 - 10.1145/3243734.3243750
DO - 10.1145/3243734.3243750
M3 - Conference contribution
AN - SCOPUS:85056896227
T3 - Proceedings of the ACM Conference on Computer and Communications Security
SP - 230
EP - 246
BT - CCS 2018 - Proceedings of the 2018 ACM SIGSAC Conference on Computer and Communications Security
PB - Association for Computing Machinery
Y2 - 15 October 2018
ER -