TY - GEN
T1 - PD-ML-Lite
T2 - 22nd International Conference on Information Security
AU - Tsikhanovich, Maksim
AU - Magdon-Ismail, Malik
AU - Ishaq, Muhammad
AU - Zikas, Vassilis
PY - 2019/9/2
Y1 - 2019/9/2
N2 - Privacy arises to a major issue in distributed learning. Current approaches that do not use a trusted external authority either reduce the accuracy of the learning algorithm (e.g., by adding noise), or incur a high performance penalty. We propose a methodology for private distributed ML from light-weight cryptography (in short, PD-ML-Lite). We apply our methodology to two major ML algorithms, namely non-negative matrix factorization (NMF) and singular value decomposition (SVD). Our protocols are communication optimal, achieve the same accuracy as their non-private counterparts, and satisfy a notion of privacy—which we define—that is both intuitive and measurable. We use light cryptographic tools (multi-party secure sum and normed secure sum) to build learning algorithms rather than wrap complex learning algorithms in a heavy multi-party computation (MPC) framework. We showcase our algorithms’ utility and privacy for NMF on topic modeling and recommender systems, and for SVD on principal component regression, and low rank approximation.
AB - Privacy arises to a major issue in distributed learning. Current approaches that do not use a trusted external authority either reduce the accuracy of the learning algorithm (e.g., by adding noise), or incur a high performance penalty. We propose a methodology for private distributed ML from light-weight cryptography (in short, PD-ML-Lite). We apply our methodology to two major ML algorithms, namely non-negative matrix factorization (NMF) and singular value decomposition (SVD). Our protocols are communication optimal, achieve the same accuracy as their non-private counterparts, and satisfy a notion of privacy—which we define—that is both intuitive and measurable. We use light cryptographic tools (multi-party secure sum and normed secure sum) to build learning algorithms rather than wrap complex learning algorithms in a heavy multi-party computation (MPC) framework. We showcase our algorithms’ utility and privacy for NMF on topic modeling and recommender systems, and for SVD on principal component regression, and low rank approximation.
UR - http://www.scopus.com/inward/record.url?scp=85072862143&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-30215-3_8
DO - 10.1007/978-3-030-30215-3_8
M3 - Conference contribution
AN - SCOPUS:85072862143
SN - 978-3-030-30214-6
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 149
EP - 167
BT - Information Security
A2 - Lin, Zhiqiang
A2 - Papamanthou, Charalampos
A2 - Polychronakis, Michalis
PB - Springer
CY - Cham
Y2 - 16 September 2019 through 18 September 2019
ER -