TY - GEN
T1 - Differentially private distributed principal component analysis
AU - Imtiaz, Hafiz
AU - Sarwate, Anand D.
N1 - Funding Information:
The work of the authors was supported by the NSF under award CCF-1453432, by the NIH under award 1R01DA040487-01A1, and by DARPA and SSC Pacific under contract No. N66001-15-C-4070.
PY - 2018/9/10
Y1 - 2018/9/10
N2 - Differential privacy is a cryptographically-motivated formal privacy definition that is robust against strong adversaries. The principal component analysis (PCA) algorithm is frequently used in signal processing, machine learning, and statistics pipelines. In many scenarios, private or sensitive data is distributed across different sites: in this paper we propose a differentially private distributed PCA scheme to enable collaborative dimensionality reduction. We investigate the performance of the proposed algorithm on synthetic and real datasets and show empirically that our algorithm can reach the same level of utility as the non-private PCA for some parameter choices, which indicates that it is possible to have meaningful utility while preserving privacy.
AB - Differential privacy is a cryptographically-motivated formal privacy definition that is robust against strong adversaries. The principal component analysis (PCA) algorithm is frequently used in signal processing, machine learning, and statistics pipelines. In many scenarios, private or sensitive data is distributed across different sites: in this paper we propose a differentially private distributed PCA scheme to enable collaborative dimensionality reduction. We investigate the performance of the proposed algorithm on synthetic and real datasets and show empirically that our algorithm can reach the same level of utility as the non-private PCA for some parameter choices, which indicates that it is possible to have meaningful utility while preserving privacy.
KW - Differential privacy
KW - Distributed algorithm
KW - Principal component analysis
UR - http://www.scopus.com/inward/record.url?scp=85054262191&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85054262191&partnerID=8YFLogxK
U2 - 10.1109/ICASSP.2018.8462519
DO - 10.1109/ICASSP.2018.8462519
M3 - Conference contribution
AN - SCOPUS:85054262191
SN - 9781538646588
T3 - ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
SP - 2206
EP - 2210
BT - 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018 - Proceedings
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2018 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2018
Y2 - 15 April 2018 through 20 April 2018
ER -