Federated singular value decomposition for high-dimensional data

被引:4
作者
Hartebrodt, Anne [1 ,2 ]
Rottger, Richard [1 ]
Blumenthal, David B. [2 ]
机构
[1] Univ Southern Denmark, Dept Math & Comp Sci, Campusvej 55, DK-5230 Odense, Denmark
[2] Friedrich Alexander Univ Erlangen Nurnberg FAU, Dept Artificial Intelligence Biomed Engn AIBE, Werner von Siemens Str 61, D-91052 Erlangen, Germany
关键词
Singular value decomposition; Federated learning; Principal component analysis; Genome-wide association studies; PRINCIPAL-COMPONENT ANALYSIS; ALGORITHMS;
D O I
10.1007/s10618-023-00983-z
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Federated learning (FL) is emerging as a privacy-aware alternative to classical cloud-based machine learning. In FL, the sensitive data remains in data silos and only aggregated parameters are exchanged. Hospitals and research institutions which are not willing to share their data can join a federated study without breaching confidentiality. In addition to the extreme sensitivity of biomedical data, the high dimensionality poses a challenge in the context of federated genome-wide association studies (GWAS). In this article, we present a federated singular value decomposition algorithm, suitable for the privacy-related and computational requirements of GWAS. Notably, the algorithm has a transmission cost independent of the number of samples and is only weakly dependent on the number of features, because the singular vectors corresponding to the samples are never exchanged and the vectors associated with the features are only transmitted to an aggregator for a fixed number of iterations. Although motivated by GWAS, the algorithm is generically applicable for both horizontally and vertically partitioned data.
引用
收藏
页码:938 / 975
页数:38
相关论文
共 61 条
[11]   Secure genome-wide association analysis using multiparty computation [J].
Cho, Hyunghoon ;
Wu, David J. ;
Berger, Bonnie .
NATURE BIOTECHNOLOGY, 2018, 36 (06) :547-+
[12]  
Cramer R., 2015, Secure Multiparty Computation and Secret Sharing, DOI [10.1017/CBO9781107337756, DOI 10.1017/CBO9781107337756]
[13]  
Dua D, 2017, UCI machine learning repository
[14]   The Algorithmic Foundations of Differential Privacy [J].
Dwork, Cynthia ;
Roth, Aaron .
FOUNDATIONS AND TRENDS IN THEORETICAL COMPUTER SCIENCE, 2013, 9 (3-4) :211-406
[15]  
Froelicher D, 2023, ARXIV
[16]   Fast Principal-Component Analysis Reveals Convergent Evolution of ADH1B in Europe and East Asia [J].
Galinsky, Kevin J. ;
Bhatia, Gaurav ;
Loh, Po-Ru ;
Georgiev, Stoyan ;
Mukherjee, Sayan ;
Patterson, Nick J. ;
Price, Alkes L. .
AMERICAN JOURNAL OF HUMAN GENETICS, 2016, 98 (03) :456-472
[17]   Consequences of PCA graphs, SNP codings, and PCA variants for elucidating population structure [J].
Gauch, Hugh G., Jr. ;
Qian, Sheng ;
Piepho, Hans-Peter ;
Zhou, Linda ;
Chen, Rui .
PLOS ONE, 2019, 14 (06)
[18]  
Gonen A., 2018, J MACH LEARN RES, V1, P1
[19]  
Grammenos Andreas, 2020, ADV NEURAL INF PROCE, V33, P6453
[20]   A covariance-free iterative algorithm for distributed principal component analysis on vertically partitioned data [J].
Guo, Yue-Fei ;
Lin, Xiaodong ;
Teng, Zhou ;
Xue, Xiangyang ;
Fan, Jianping .
PATTERN RECOGNITION, 2012, 45 (03) :1211-1219