Stochastic privacy-preserving methods for nonconvex sparse learning

Sparse learning is essential in mining high-dimensional data. Iterative hard thresholding (IHT) methods are effective for optimizing nonconvex objectives for sparse learning. However, IHT methods are vulnerable to adversary attacks that infer sensitive data. Although pioneering works attempted to re...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Information sciences Ročník 630; s. 567 - 585
Hlavní autoři: Liang, Guannan, Tong, Qianqian, Ding, Jiahao, Pan, Miao, Bi, Jinbo
Médium: Journal Article
Jazyk:angličtina
Vydáno: Elsevier Inc 01.06.2023
Témata:
ISSN:0020-0255, 1872-6291
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Abstract Sparse learning is essential in mining high-dimensional data. Iterative hard thresholding (IHT) methods are effective for optimizing nonconvex objectives for sparse learning. However, IHT methods are vulnerable to adversary attacks that infer sensitive data. Although pioneering works attempted to relieve such vulnerability, they confront the issue of high computational cost for large-scale problems. We propose two differentially private stochastic IHT: one based on the stochastic gradient descent method (DP-SGD-HT) and the other based on the stochastically controlled stochastic gradient method (DP-SCSG-HT). The DP-SGD-HT method perturbs stochastic gradients with small Gaussian noise rather than full gradients, which are computationally expensive. As a result, computational complexity is reduced from O(nlog(n)) to a lower O(blog(n)), where n is the sample size and b is the mini-batch size used to compute stochastic gradients. The DP-SCSG-HT method further perturbs the stochastic gradients controlled by large-batch snapshot gradients to reduce stochastic gradient variance. We prove that both algorithms guarantee differential privacy and have linear convergence rates with estimation bias. A utility analysis examines the relationship between convergence rate and the level of perturbation, yielding the best-known utility bound for nonconvex sparse optimization. Extensive experiments show that our algorithms outperform existing methods.
AbstractList Sparse learning is essential in mining high-dimensional data. Iterative hard thresholding (IHT) methods are effective for optimizing nonconvex objectives for sparse learning. However, IHT methods are vulnerable to adversary attacks that infer sensitive data. Although pioneering works attempted to relieve such vulnerability, they confront the issue of high computational cost for large-scale problems. We propose two differentially private stochastic IHT: one based on the stochastic gradient descent method (DP-SGD-HT) and the other based on the stochastically controlled stochastic gradient method (DP-SCSG-HT). The DP-SGD-HT method perturbs stochastic gradients with small Gaussian noise rather than full gradients, which are computationally expensive. As a result, computational complexity is reduced from O(nlog(n)) to a lower O(blog(n)), where n is the sample size and b is the mini-batch size used to compute stochastic gradients. The DP-SCSG-HT method further perturbs the stochastic gradients controlled by large-batch snapshot gradients to reduce stochastic gradient variance. We prove that both algorithms guarantee differential privacy and have linear convergence rates with estimation bias. A utility analysis examines the relationship between convergence rate and the level of perturbation, yielding the best-known utility bound for nonconvex sparse optimization. Extensive experiments show that our algorithms outperform existing methods.
Author Tong, Qianqian
Bi, Jinbo
Liang, Guannan
Pan, Miao
Ding, Jiahao
Author_xml – sequence: 1
  givenname: Guannan
  surname: Liang
  fullname: Liang, Guannan
  email: guannan.liang@uconn.edu
  organization: University of Connecticut, United States
– sequence: 2
  givenname: Qianqian
  surname: Tong
  fullname: Tong, Qianqian
  email: qianqian.tong@uconn.edu
  organization: University of Connecticut, United States
– sequence: 3
  givenname: Jiahao
  surname: Ding
  fullname: Ding, Jiahao
  email: jding7@uh.edu
  organization: University of Houston, United States
– sequence: 4
  givenname: Miao
  surname: Pan
  fullname: Pan, Miao
  email: mpan2@uh.edu
  organization: University of Houston, United States
– sequence: 5
  givenname: Jinbo
  surname: Bi
  fullname: Bi, Jinbo
  email: jinbo.bi@uconn.edu
  organization: University of Connecticut, United States
BookMark eNp9kEFOwzAQRS1UJNrCAdjlAgljJ7FjsUIVUKRKLIC15Thj6qq1KzuK6O1xVdas_ua_ma-3IDMfPBJyT6GiQPnDrnI-VQwYq0BWwNkVmdNOsJIzSWdkDsCgBNa2N2SR0g4AGsH5nKw-xmC2Oo3OFMfoJm1O5TFiwjg5_10ccNyGIRU2xCJ_NMFP-FOko44Jiz3q6HPrllxbvU9495dL8vXy_Llal5v317fV06Y0rJFj2VjNBLYoB9sKawXX2nCLBrGrZdcL4GhtbQ0ILinvjbSyMT0C7Rl2XU3rJaGXuyaGlCJalRcfdDwpCupsQe1UtqDOFhRIlS1k5vHCYB42OYwqGYfe4OAimlENwf1D_wLHRmmI
Cites_doi 10.1109/MSP.2012.2211477
10.24963/ijcai.2019/519
10.1137/090756090
10.1016/j.cell.2018.02.010
10.1038/s41598-022-05539-7
10.1016/j.acha.2009.04.002
10.1109/ACCESS.2022.3151670
10.1016/j.ins.2022.05.128
10.1016/j.ins.2021.01.058
10.1016/j.ins.2022.04.030
10.1016/j.ins.2019.07.001
10.1109/78.258082
10.1609/aaai.v34i02.5519
10.1109/TIT.2017.2749330
ContentType Journal Article
Copyright 2022
Copyright_xml – notice: 2022
DBID AAYXX
CITATION
DOI 10.1016/j.ins.2022.09.062
DatabaseName CrossRef
DatabaseTitle CrossRef
DatabaseTitleList
DeliveryMethod fulltext_linktorsrc
Discipline Engineering
Library & Information Science
EISSN 1872-6291
EndPage 585
ExternalDocumentID 10_1016_j_ins_2022_09_062
S0020025522011161
GroupedDBID --K
--M
--Z
-~X
.DC
.~1
0R~
1B1
1OL
1RT
1~.
1~5
29I
4.4
457
4G.
5GY
5VS
7-5
71M
8P~
9JN
9JO
AAAKF
AAAKG
AABNK
AACTN
AAEDT
AAEDW
AAIAV
AAIKJ
AAKOC
AALRI
AAOAW
AAQFI
AAQXK
AARIN
AAXUO
AAYFN
ABAOU
ABBOA
ABEFU
ABFNM
ABJNI
ABMAC
ABTAH
ABUCO
ABXDB
ABYKQ
ACAZW
ACDAQ
ACGFS
ACNNM
ACRLP
ACZNC
ADBBV
ADEZE
ADGUI
ADJOM
ADMUD
ADTZH
AEBSH
AECPX
AEKER
AENEX
AFFNX
AFKWA
AFTJW
AGHFR
AGUBO
AGYEJ
AHHHB
AHJVU
AHZHX
AIALX
AIEXJ
AIGVJ
AIKHN
AITUG
AJBFU
AJOXV
ALMA_UNASSIGNED_HOLDINGS
AMFUW
AMRAJ
AOUOD
APLSM
ARUGR
ASPBG
AVWKF
AXJTR
AZFZN
BJAXD
BKOJK
BLXMC
CS3
DU5
EBS
EFJIC
EFLBG
EJD
EO8
EO9
EP2
EP3
F5P
FDB
FEDTE
FGOYB
FIRID
FNPLU
FYGXN
G-Q
GBLVA
GBOLZ
HAMUX
HLZ
HVGLF
HZ~
H~9
IHE
J1W
JJJVA
KOM
LG9
LY1
M41
MHUIS
MO0
MS~
N9A
O-L
O9-
OAUVE
OZT
P-8
P-9
P2P
PC.
Q38
R2-
RIG
ROL
RPZ
SBC
SDF
SDG
SDP
SDS
SES
SEW
SPC
SPCBC
SSB
SSD
SST
SSV
SSW
SSZ
T5K
TN5
TWZ
UHS
WH7
WUQ
XPP
YYP
ZMT
ZY4
~02
~G-
77I
9DU
AATTM
AAXKI
AAYWO
AAYXX
ABWVN
ACLOT
ACRPL
ACVFH
ADCNI
ADNMO
ADVLN
AEIPS
AEUPX
AFJKZ
AFPUW
AGQPQ
AIGII
AIIUN
AKBMS
AKRWK
AKYEP
ANKPU
APXCP
CITATION
EFKBS
~HD
ID FETCH-LOGICAL-c249t-4fa27e5e9df57ff76aac6fecee8398b706eff3fc076916bc9f94cbe01b2e88313
ISICitedReferencesCount 1
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000946723000001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN 0020-0255
IngestDate Sat Nov 29 07:25:32 EST 2025
Fri Feb 23 02:37:44 EST 2024
IsPeerReviewed true
IsScholarly true
Keywords Differential privacy
Stochastic algorithm
Sparse learning
Language English
LinkModel OpenURL
MergedId FETCHMERGED-LOGICAL-c249t-4fa27e5e9df57ff76aac6fecee8398b706eff3fc076916bc9f94cbe01b2e88313
PageCount 19
ParticipantIDs crossref_primary_10_1016_j_ins_2022_09_062
elsevier_sciencedirect_doi_10_1016_j_ins_2022_09_062
PublicationCentury 2000
PublicationDate June 2023
2023-06-00
PublicationDateYYYYMMDD 2023-06-01
PublicationDate_xml – month: 06
  year: 2023
  text: June 2023
PublicationDecade 2020
PublicationTitle Information sciences
PublicationYear 2023
Publisher Elsevier Inc
Publisher_xml – name: Elsevier Inc
References El Ouadrhiri, Abdelhadi (b0065) 2022; 10
X. Li, R. Arora, H. Liu, J. Haupt, and T. Zhao. Nonconvex sparse learning via stochastic optimization with progressive variance reduction. arXiv preprint arXiv:1605.02711, 2016a.
Jiang, Niu, Ying, Wu, Luo (b0095) 2022; 602
Li, Li, Zhu, Huang (b0135) 2019; 503
Nguyen, Needell, Woolf (b0165) 2017; 63
A. Rényi et al. On measures of entropy and information. In Proceedings of the Fourth Berkeley Symposium on Mathematical Statistics and Probability, Volume 1: Contributions to the Theory of Statistics. The Regents of the University of California, 1961.
P. Jain, A. Tewari, and P. Kar. On iterative hard thresholding methods for high-dimensional m-estimation. In Advances in Neural Information Processing Systems, pages 685–693, 2014.
P. Zhou, X. Yuan, and J. Feng. Efficient stochastic gradient hard thresholding. In Advances in Neural Information Processing Systems, pages 1988–1997, 2018.
M. Elibol, L. Lei, and M.I. Jordan. Variance reduction with sparse gradients. arXiv preprint arXiv:2001.09623, 2020.
Guo, Ding, Wang, Jia (b0080) 2022
Kermany, Goldbaum, Cai, Valentim, Liang, Baxter, McKeown, Yang, Wu, Yan (b0110) 2018; 172
Pati, Rezaiifar, Krishnaprasad (b0170) 1993
Wang, Wang (b0205) 2021; 560
L. Wang and Q. Gu. A knowledge transfer framework for differentially private sparse learning. arXiv preprint arXiv:1909.06322, 2019b.
Erlingsson, Pihur, Korolova (b0075) 2014
T. Li, A.K. Sahu, M. Zaheer, M. Sanjabi, A. Talwalkar, and V. Smith. Federated optimization in heterogeneous networks. arXiv preprint arXiv:1812.06127, 2018.
K. Talwar, A.G. Thakurta, and L. Zhang. Nearly optimal private lasso. In Advances in Neural Information Processing Systems, pages 3025–3033, 2015.
Wang, Xu (b0195) 2019
Mironov (b0155) 2017
M. Abadi, A. Agarwal, P. Barham, E. Brevdo, Z. Chen, C. Citro, G.S. Corrado, A. Davis, J. Dean, M. Devin, et al. Tensorflow: Large-scale machine learning on heterogeneous distributed systems. arXiv preprint arXiv:1603.04467, 2016a.
D. Kifer, A. Smith, and A. Thakurta. Private convex empirical risk minimization and high-dimensional regression. In Conference on Learning Theory, pages 25–1, 2012.
Dwork, Rothblum, Vadhan (b0060) 2010
Dwork, Kenthapadi, McSherry, Mironov, Naor (b0050) 2006
Q. Tong, G. Liang, T. Zhu, and J. Bi. Federated nonconvex sparse learning. arXiv preprint arXiv:2101.00052, 2020.
J. Near. Differential privacy at scale: Uber and berkeley collaboration. In Enigma 2018 (Enigma 2018), 2018.
R. Johnson and T. Zhang. Accelerating stochastic gradient descent using predictive variance reduction. In Advances in neural information processing systems, pages 315–323, 2013.
D. Wang, M. Ye, and J. Xu. Differentially private empirical risk minimization revisited: Faster and more general. In Advances in Neural Information Processing Systems, pages 2722–2731, 2017.
Wu, Li, Kumar, Chaudhuri, Jha, Naughton (b0230) 2017
Kasiviswanathan, Lee, Nissim, Raskhodnikova, Smith (b0105) 2011; 40
A.E.C. Cloud. Amazon web services. Retrieved November, 9 (2011): 2011, 2011.
L. Lei and M. Jordan. Less than a single pass: Stochastically controlled stochastic gradient. In Artificial Intelligence and Statistics, pages 148–156, 2017.
Deng (b0045) 2012; 29
Dwork, McSherry, Nissim, Smith (b0055) 2006
L. Wang and Q. Gu. Differentially private iterative gradient hard thresholding for sparse learning. In 28th International Joint Conference on Artificial Intelligence, 2019a.
Y.-X. Wang, B. Balle, and S. Kasiviswanathan. Subsampled r⧹)ényi differential privacy and analytical moments accountant. arXiv preprint arXiv:1808.00087, 2018.
Zhu, Fang, Guo, Niu, Cao, Yue, Liu (b0240) 2009
Abadi, Chu, Goodfellow, McMahan, Mironov, Talwar, Zhang (b0010) 2016
Adnan, Kalra, Cresswell, Taylor, Tizhoosh (b0015) 2022; 12
L. Wang, B. Jayaraman, D. Evans, and Q. Gu. Efficient privacy-preserving nonconvex optimization. arXiv preprint arXiv:1910.13659, 2019.
Bahmani, Raj, Boufounos (b0020) 2013; 14
Bassily, Smith, Thakurta (b0025) 2014
Truex, Liu, Gursoy, Yu, Wei (b0190) 2019
X. Li, T. Zhao, R. Arora, H. Liu, and J. Haupt. Stochastic variance reduced optimization for nonconvex sparse learning. In International Conference on Machine Learning, pages 917–925, 2016b.
G. Liang, Q. Tong, C.J. Zhu, and J. Bi. An effective hard thresholding method based on stochastic variance reduction for nonconvex sparse learning. In AAAI, pages 1585–1592, 2020.
Chaudhuri, Monteleoni, Sarwate (b0035) 2011; 12
Blumensath, Davies (b0030) 2009; 27
R.B. Harikandeh, M.O. Ahmed, A. Virani, M. Schmidt, J. Konečný, and S. Sallinen. Stopwasting my gradients: Practical svrg. In Advances in Neural Information Processing Systems, pages 2251–2259, 2015.
Mallat, Zhang (b0150) 1993; 41
10.1016/j.ins.2022.09.062_b0120
El Ouadrhiri (10.1016/j.ins.2022.09.062_b0065) 2022; 10
10.1016/j.ins.2022.09.062_b0185
10.1016/j.ins.2022.09.062_b0100
10.1016/j.ins.2022.09.062_b0220
10.1016/j.ins.2022.09.062_b0145
10.1016/j.ins.2022.09.062_b0200
10.1016/j.ins.2022.09.062_b0005
Jiang (10.1016/j.ins.2022.09.062_b0095) 2022; 602
10.1016/j.ins.2022.09.062_b0225
10.1016/j.ins.2022.09.062_b0125
Kermany (10.1016/j.ins.2022.09.062_b0110) 2018; 172
10.1016/j.ins.2022.09.062_b0180
10.1016/j.ins.2022.09.062_b0160
10.1016/j.ins.2022.09.062_b0085
10.1016/j.ins.2022.09.062_b0140
10.1016/j.ins.2022.09.062_b0040
Blumensath (10.1016/j.ins.2022.09.062_b0030) 2009; 27
Chaudhuri (10.1016/j.ins.2022.09.062_b0035) 2011; 12
Mironov (10.1016/j.ins.2022.09.062_b0155) 2017
Truex (10.1016/j.ins.2022.09.062_b0190) 2019
Wang (10.1016/j.ins.2022.09.062_b0205) 2021; 560
Abadi (10.1016/j.ins.2022.09.062_b0010) 2016
Deng (10.1016/j.ins.2022.09.062_b0045) 2012; 29
Bassily (10.1016/j.ins.2022.09.062_b0025) 2014
Dwork (10.1016/j.ins.2022.09.062_b0050) 2006
Wu (10.1016/j.ins.2022.09.062_b0230) 2017
10.1016/j.ins.2022.09.062_b0215
Dwork (10.1016/j.ins.2022.09.062_b0055) 2006
10.1016/j.ins.2022.09.062_b0175
10.1016/j.ins.2022.09.062_b0130
10.1016/j.ins.2022.09.062_b0210
10.1016/j.ins.2022.09.062_b0115
Guo (10.1016/j.ins.2022.09.062_b0080) 2022
Li (10.1016/j.ins.2022.09.062_b0135) 2019; 503
10.1016/j.ins.2022.09.062_b0235
10.1016/j.ins.2022.09.062_b0090
10.1016/j.ins.2022.09.062_b0070
Mallat (10.1016/j.ins.2022.09.062_b0150) 1993; 41
Dwork (10.1016/j.ins.2022.09.062_b0060) 2010
Bahmani (10.1016/j.ins.2022.09.062_b0020) 2013; 14
Zhu (10.1016/j.ins.2022.09.062_b0240) 2009
Adnan (10.1016/j.ins.2022.09.062_b0015) 2022; 12
Kasiviswanathan (10.1016/j.ins.2022.09.062_b0105) 2011; 40
Wang (10.1016/j.ins.2022.09.062_b0195) 2019
Nguyen (10.1016/j.ins.2022.09.062_b0165) 2017; 63
Pati (10.1016/j.ins.2022.09.062_b0170) 1993
Erlingsson (10.1016/j.ins.2022.09.062_b0075) 2014
References_xml – reference: T. Li, A.K. Sahu, M. Zaheer, M. Sanjabi, A. Talwalkar, and V. Smith. Federated optimization in heterogeneous networks. arXiv preprint arXiv:1812.06127, 2018.
– reference: P. Zhou, X. Yuan, and J. Feng. Efficient stochastic gradient hard thresholding. In Advances in Neural Information Processing Systems, pages 1988–1997, 2018.
– volume: 172
  start-page: 1122
  year: 2018
  end-page: 1131
  ident: b0110
  article-title: Identifying medical diagnoses and treatable diseases by image-based deep learning
  publication-title: Cell
– reference: R.B. Harikandeh, M.O. Ahmed, A. Virani, M. Schmidt, J. Konečný, and S. Sallinen. Stopwasting my gradients: Practical svrg. In Advances in Neural Information Processing Systems, pages 2251–2259, 2015.
– volume: 503
  start-page: 219
  year: 2019
  end-page: 237
  ident: b0135
  article-title: The optimal upper bound of the number of queries for laplace mechanism under differential privacy
  publication-title: Information Sciences
– reference: A. Rényi et al. On measures of entropy and information. In Proceedings of the Fourth Berkeley Symposium on Mathematical Statistics and Probability, Volume 1: Contributions to the Theory of Statistics. The Regents of the University of California, 1961.
– start-page: 308
  year: 2016
  end-page: 318
  ident: b0010
  article-title: Deep learning with differential privacy
  publication-title: Proceedings of the 2016 ACM SIGSAC Conference on Computer and Communications Security
– reference: L. Wang and Q. Gu. A knowledge transfer framework for differentially private sparse learning. arXiv preprint arXiv:1909.06322, 2019b.
– volume: 602
  start-page: 57
  year: 2022
  end-page: 74
  ident: b0095
  article-title: Pricing gan-based data generators under rényi differential privacy
  publication-title: Information Sciences
– reference: L. Wang and Q. Gu. Differentially private iterative gradient hard thresholding for sparse learning. In 28th International Joint Conference on Artificial Intelligence, 2019a.
– volume: 29
  start-page: 141
  year: 2012
  end-page: 142
  ident: b0045
  article-title: The mnist database of handwritten digit images for machine learning research [best of the web]
  publication-title: IEEE Signal Processing Magazine
– reference: J. Near. Differential privacy at scale: Uber and berkeley collaboration. In Enigma 2018 (Enigma 2018), 2018.
– volume: 40
  start-page: 793
  year: 2011
  end-page: 826
  ident: b0105
  article-title: What can we learn privately?
  publication-title: SIAM Journal on Computing
– reference: L. Wang, B. Jayaraman, D. Evans, and Q. Gu. Efficient privacy-preserving nonconvex optimization. arXiv preprint arXiv:1910.13659, 2019.
– volume: 12
  start-page: 1
  year: 2022
  end-page: 10
  ident: b0015
  article-title: Federated learning and differential privacy for medical image analysis
  publication-title: Scientific reports
– reference: X. Li, R. Arora, H. Liu, J. Haupt, and T. Zhao. Nonconvex sparse learning via stochastic optimization with progressive variance reduction. arXiv preprint arXiv:1605.02711, 2016a.
– volume: 14
  start-page: 807
  year: 2013
  end-page: 841
  ident: b0020
  article-title: Greedy sparsity-constrained optimization
  publication-title: Journal of Machine Learning Research
– year: 2022
  ident: b0080
  article-title: Combinatorial resources auction in decentralized edge-thing systems using blockchain and differential privacy
  publication-title: Information Sciences
– start-page: 486
  year: 2006
  end-page: 503
  ident: b0050
  article-title: Our data, ourselves: Privacy via distributed noise generation
  publication-title: Annual International Conference on the Theory and Applications of Cryptographic Techniques
– start-page: 6628
  year: 2019
  end-page: 6637
  ident: b0195
  article-title: On sparse linear regression in the local differential privacy model
  publication-title: International Conference on Machine Learning
– reference: D. Wang, M. Ye, and J. Xu. Differentially private empirical risk minimization revisited: Faster and more general. In Advances in Neural Information Processing Systems, pages 2722–2731, 2017.
– reference: P. Jain, A. Tewari, and P. Kar. On iterative hard thresholding methods for high-dimensional m-estimation. In Advances in Neural Information Processing Systems, pages 685–693, 2014.
– volume: 10
  start-page: 22359
  year: 2022
  end-page: 22380
  ident: b0065
  article-title: Differential privacy for deep and federated learning: A survey
  publication-title: IEEE Access
– reference: X. Li, T. Zhao, R. Arora, H. Liu, and J. Haupt. Stochastic variance reduced optimization for nonconvex sparse learning. In International Conference on Machine Learning, pages 917–925, 2016b.
– reference: Q. Tong, G. Liang, T. Zhu, and J. Bi. Federated nonconvex sparse learning. arXiv preprint arXiv:2101.00052, 2020.
– reference: D. Kifer, A. Smith, and A. Thakurta. Private convex empirical risk minimization and high-dimensional regression. In Conference on Learning Theory, pages 25–1, 2012.
– volume: 12
  start-page: 1069
  year: 2011
  end-page: 1109
  ident: b0035
  article-title: Differentially private empirical risk minimization
  publication-title: Journal of Machine Learning Research
– reference: G. Liang, Q. Tong, C.J. Zhu, and J. Bi. An effective hard thresholding method based on stochastic variance reduction for nonconvex sparse learning. In AAAI, pages 1585–1592, 2020.
– volume: 560
  start-page: 347
  year: 2021
  end-page: 369
  ident: b0205
  article-title: Correlated tuple data release via differential privacy
  publication-title: Information Sciences
– start-page: 1307
  year: 2017
  end-page: 1322
  ident: b0230
  article-title: Bolt-on differential privacy for scalable stochastic gradient descent-based analytics
  publication-title: Proceedings of the 2017 ACM International Conference on Management of Data
– reference: M. Abadi, A. Agarwal, P. Barham, E. Brevdo, Z. Chen, C. Citro, G.S. Corrado, A. Davis, J. Dean, M. Devin, et al. Tensorflow: Large-scale machine learning on heterogeneous distributed systems. arXiv preprint arXiv:1603.04467, 2016a.
– reference: A.E.C. Cloud. Amazon web services. Retrieved November, 9 (2011): 2011, 2011.
– volume: 63
  start-page: 6869
  year: 2017
  end-page: 6895
  ident: b0165
  article-title: Linear convergence of stochastic iterative greedy algorithms with sparse constraints
  publication-title: IEEE Transactions on Information Theory
– start-page: 265
  year: 2006
  end-page: 284
  ident: b0055
  article-title: Calibrating noise to sensitivity in private data analysis
  publication-title: Theory of cryptography conference
– start-page: 1054
  year: 2014
  end-page: 1067
  ident: b0075
  article-title: Rappor: Randomized aggregatable privacy-preserving ordinal response
  publication-title: Proceedings of the 2014 ACM SIGSAC conference on computer and communications security
– reference: L. Lei and M. Jordan. Less than a single pass: Stochastically controlled stochastic gradient. In Artificial Intelligence and Statistics, pages 148–156, 2017.
– volume: 27
  start-page: 265
  year: 2009
  end-page: 274
  ident: b0030
  article-title: Iterative hard thresholding for compressed sensing
  publication-title: Applied and computational harmonic analysis
– start-page: 51
  year: 2010
  end-page: 60
  ident: b0060
  article-title: Boosting and differential privacy
  publication-title: 2010 IEEE 51st Annual Symposium on Foundations of Computer Science
– reference: M. Elibol, L. Lei, and M.I. Jordan. Variance reduction with sparse gradients. arXiv preprint arXiv:2001.09623, 2020.
– volume: 41
  start-page: 3397
  year: 1993
  end-page: 3415
  ident: b0150
  article-title: Matching pursuits with time-frequency dictionaries
  publication-title: IEEE Transactions on signal processing
– reference: Y.-X. Wang, B. Balle, and S. Kasiviswanathan. Subsampled r⧹)ényi differential privacy and analytical moments accountant. arXiv preprint arXiv:1808.00087, 2018.
– reference: K. Talwar, A.G. Thakurta, and L. Zhang. Nearly optimal private lasso. In Advances in Neural Information Processing Systems, pages 3025–3033, 2015.
– reference: R. Johnson and T. Zhang. Accelerating stochastic gradient descent using predictive variance reduction. In Advances in neural information processing systems, pages 315–323, 2013.
– start-page: 464
  year: 2014
  end-page: 473
  ident: b0025
  article-title: Private empirical risk minimization: Efficient algorithms and tight error bounds
  publication-title: 2014 IEEE 55th Annual Symposium on Foundations of Computer Science
– year: 2019
  ident: b0190
  article-title: Demystifying membership inference attacks in machine learning as a service
  publication-title: IEEE Transactions on Services Computing
– start-page: 263
  year: 2017
  end-page: 275
  ident: b0155
  article-title: Rényi differential privacy
  publication-title: 2017 IEEE 30th Computer Security Foundations Symposium (CSF)
– start-page: 40
  year: 1993
  end-page: 44
  ident: b0170
  article-title: Orthogonal matching pursuit: Recursive function approximation with applications to wavelet decomposition
  publication-title: Proceedings of 27th Asilomar conference on signals, systems and computers
– start-page: 621
  year: 2009
  end-page: 625
  ident: b0240
  article-title: Ibm cloud computing powering a smarter planet
  publication-title: IEEE International Conference on Cloud Computing
– volume: 29
  start-page: 141
  issue: 6
  year: 2012
  ident: 10.1016/j.ins.2022.09.062_b0045
  article-title: The mnist database of handwritten digit images for machine learning research [best of the web]
  publication-title: IEEE Signal Processing Magazine
  doi: 10.1109/MSP.2012.2211477
– ident: 10.1016/j.ins.2022.09.062_b0180
– ident: 10.1016/j.ins.2022.09.062_b0235
– ident: 10.1016/j.ins.2022.09.062_b0130
– ident: 10.1016/j.ins.2022.09.062_b0210
  doi: 10.24963/ijcai.2019/519
– volume: 40
  start-page: 793
  issue: 3
  year: 2011
  ident: 10.1016/j.ins.2022.09.062_b0105
  article-title: What can we learn privately?
  publication-title: SIAM Journal on Computing
  doi: 10.1137/090756090
– ident: 10.1016/j.ins.2022.09.062_b0040
– start-page: 486
  year: 2006
  ident: 10.1016/j.ins.2022.09.062_b0050
  article-title: Our data, ourselves: Privacy via distributed noise generation
– volume: 172
  start-page: 1122
  issue: 5
  year: 2018
  ident: 10.1016/j.ins.2022.09.062_b0110
  article-title: Identifying medical diagnoses and treatable diseases by image-based deep learning
  publication-title: Cell
  doi: 10.1016/j.cell.2018.02.010
– ident: 10.1016/j.ins.2022.09.062_b0120
– ident: 10.1016/j.ins.2022.09.062_b0225
– start-page: 621
  year: 2009
  ident: 10.1016/j.ins.2022.09.062_b0240
  article-title: Ibm cloud computing powering a smarter planet
– ident: 10.1016/j.ins.2022.09.062_b0070
– ident: 10.1016/j.ins.2022.09.062_b0200
– ident: 10.1016/j.ins.2022.09.062_b0185
– volume: 12
  start-page: 1
  issue: 1
  year: 2022
  ident: 10.1016/j.ins.2022.09.062_b0015
  article-title: Federated learning and differential privacy for medical image analysis
  publication-title: Scientific reports
  doi: 10.1038/s41598-022-05539-7
– volume: 27
  start-page: 265
  issue: 3
  year: 2009
  ident: 10.1016/j.ins.2022.09.062_b0030
  article-title: Iterative hard thresholding for compressed sensing
  publication-title: Applied and computational harmonic analysis
  doi: 10.1016/j.acha.2009.04.002
– volume: 10
  start-page: 22359
  year: 2022
  ident: 10.1016/j.ins.2022.09.062_b0065
  article-title: Differential privacy for deep and federated learning: A survey
  publication-title: IEEE Access
  doi: 10.1109/ACCESS.2022.3151670
– start-page: 1054
  year: 2014
  ident: 10.1016/j.ins.2022.09.062_b0075
  article-title: Rappor: Randomized aggregatable privacy-preserving ordinal response
– year: 2022
  ident: 10.1016/j.ins.2022.09.062_b0080
  article-title: Combinatorial resources auction in decentralized edge-thing systems using blockchain and differential privacy
  publication-title: Information Sciences
  doi: 10.1016/j.ins.2022.05.128
– start-page: 40
  year: 1993
  ident: 10.1016/j.ins.2022.09.062_b0170
  article-title: Orthogonal matching pursuit: Recursive function approximation with applications to wavelet decomposition
– start-page: 308
  year: 2016
  ident: 10.1016/j.ins.2022.09.062_b0010
  article-title: Deep learning with differential privacy
– ident: 10.1016/j.ins.2022.09.062_b0215
– start-page: 464
  year: 2014
  ident: 10.1016/j.ins.2022.09.062_b0025
  article-title: Private empirical risk minimization: Efficient algorithms and tight error bounds
– volume: 12
  start-page: 1069
  issue: Mar
  year: 2011
  ident: 10.1016/j.ins.2022.09.062_b0035
  article-title: Differentially private empirical risk minimization
  publication-title: Journal of Machine Learning Research
– ident: 10.1016/j.ins.2022.09.062_b0160
– ident: 10.1016/j.ins.2022.09.062_b0005
– start-page: 51
  year: 2010
  ident: 10.1016/j.ins.2022.09.062_b0060
  article-title: Boosting and differential privacy
– volume: 560
  start-page: 347
  year: 2021
  ident: 10.1016/j.ins.2022.09.062_b0205
  article-title: Correlated tuple data release via differential privacy
  publication-title: Information Sciences
  doi: 10.1016/j.ins.2021.01.058
– ident: 10.1016/j.ins.2022.09.062_b0175
– start-page: 265
  year: 2006
  ident: 10.1016/j.ins.2022.09.062_b0055
  article-title: Calibrating noise to sensitivity in private data analysis
– ident: 10.1016/j.ins.2022.09.062_b0100
– start-page: 6628
  year: 2019
  ident: 10.1016/j.ins.2022.09.062_b0195
  article-title: On sparse linear regression in the local differential privacy model
– volume: 602
  start-page: 57
  year: 2022
  ident: 10.1016/j.ins.2022.09.062_b0095
  article-title: Pricing gan-based data generators under rényi differential privacy
  publication-title: Information Sciences
  doi: 10.1016/j.ins.2022.04.030
– volume: 503
  start-page: 219
  year: 2019
  ident: 10.1016/j.ins.2022.09.062_b0135
  article-title: The optimal upper bound of the number of queries for laplace mechanism under differential privacy
  publication-title: Information Sciences
  doi: 10.1016/j.ins.2019.07.001
– start-page: 1307
  year: 2017
  ident: 10.1016/j.ins.2022.09.062_b0230
  article-title: Bolt-on differential privacy for scalable stochastic gradient descent-based analytics
– year: 2019
  ident: 10.1016/j.ins.2022.09.062_b0190
  article-title: Demystifying membership inference attacks in machine learning as a service
  publication-title: IEEE Transactions on Services Computing
– volume: 41
  start-page: 3397
  issue: 12
  year: 1993
  ident: 10.1016/j.ins.2022.09.062_b0150
  article-title: Matching pursuits with time-frequency dictionaries
  publication-title: IEEE Transactions on signal processing
  doi: 10.1109/78.258082
– ident: 10.1016/j.ins.2022.09.062_b0085
– volume: 14
  start-page: 807
  issue: Mar
  year: 2013
  ident: 10.1016/j.ins.2022.09.062_b0020
  article-title: Greedy sparsity-constrained optimization
  publication-title: Journal of Machine Learning Research
– start-page: 263
  year: 2017
  ident: 10.1016/j.ins.2022.09.062_b0155
  article-title: Rényi differential privacy
– ident: 10.1016/j.ins.2022.09.062_b0115
– ident: 10.1016/j.ins.2022.09.062_b0140
– ident: 10.1016/j.ins.2022.09.062_b0125
– ident: 10.1016/j.ins.2022.09.062_b0145
  doi: 10.1609/aaai.v34i02.5519
– ident: 10.1016/j.ins.2022.09.062_b0220
– ident: 10.1016/j.ins.2022.09.062_b0090
– volume: 63
  start-page: 6869
  issue: 11
  year: 2017
  ident: 10.1016/j.ins.2022.09.062_b0165
  article-title: Linear convergence of stochastic iterative greedy algorithms with sparse constraints
  publication-title: IEEE Transactions on Information Theory
  doi: 10.1109/TIT.2017.2749330
SSID ssj0004766
Score 2.4110239
Snippet Sparse learning is essential in mining high-dimensional data. Iterative hard thresholding (IHT) methods are effective for optimizing nonconvex objectives for...
SourceID crossref
elsevier
SourceType Index Database
Publisher
StartPage 567
SubjectTerms Differential privacy
Sparse learning
Stochastic algorithm
Title Stochastic privacy-preserving methods for nonconvex sparse learning
URI https://dx.doi.org/10.1016/j.ins.2022.09.062
Volume 630
WOSCitedRecordID wos000946723000001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
journalDatabaseRights – providerCode: PRVESC
  databaseName: Elsevier SD Freedom Collection Journals 2021
  customDbUrl:
  eissn: 1872-6291
  dateEnd: 99991231
  omitProxy: false
  ssIdentifier: ssj0004766
  issn: 0020-0255
  databaseCode: AIEXJ
  dateStart: 19950101
  isFulltext: true
  titleUrlDefault: https://www.sciencedirect.com
  providerName: Elsevier
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwtV1La9tAEF6C3UN7KHmVpnmwh9JDzIKsxz6OJg-SUEJKHfBNrDa7iXNQHNsJzr_PSDt6NKmhPvQizBqNbM3HzDe78yDke5ZFlnMZsUAoxWJpFdMycyyIXN8BIQev5cphE-LyUo5G6grTbWflOAGR53KxUJP_qmpYA2UXpbMrqLsWCgvwGZQOV1A7XP9J8b_nD-ZOF-2Xe5Pp-FmbF1YkuxY2Ib_FidFlE4YeRP5l0vmiB2ZlOrPVCInbNmPFeqUSJuguaxr-c4y7zQC0PG9wNsQ831_w_WMLf8c4QOVirO_0Q3N6hQn8uIR7EGHU5ErVNQEBK6KTtl3leODiLWPip26gk038nJ539ttvJdxD0FG0Ug_Dsgctmus_emW_8WF1ZmGVtHafgoi0EJEGKg0KL90NRaJkh3QH5yeji6Z4VvgD7eovVEffZRLgm9_xd_LSIiTDdfIZIwk68AjYIGs23ySfWv0lN8k-VqXQH7SlRor2fIscNVih77FCESsU7qQ1VqjHCq2wsk2uT0-GR2cMp2owA6H2nMVOh8ImVt24RDgnuNaGOwtkCbiyzETArXORM4HgEDpkRjkVm8wG_Sy0Ukb96AvpwDPtV0KBLWugs8ZEkscC_BXXIFTeAMcH8SHfIYfV60onvnlKulRBOySuXmiKcPasLgVwLL_t2yrP2CUfG_jukc58-mT3yQfzPB_PpgeIjFfCX3zG
linkProvider Elsevier
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Stochastic+privacy-preserving+methods+for+nonconvex+sparse+learning&rft.jtitle=Information+sciences&rft.au=Liang%2C+Guannan&rft.au=Tong%2C+Qianqian&rft.au=Ding%2C+Jiahao&rft.au=Pan%2C+Miao&rft.date=2023-06-01&rft.issn=0020-0255&rft.volume=630&rft.spage=567&rft.epage=585&rft_id=info:doi/10.1016%2Fj.ins.2022.09.062&rft.externalDBID=n%2Fa&rft.externalDocID=10_1016_j_ins_2022_09_062
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=0020-0255&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=0020-0255&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=0020-0255&client=summon