TY - GEN
T1 - AN EFFICIENT HIERARCHICAL BLOCK COORDINATE DESCENT METHOD FOR TIME-VARYING GRAPHICAL LASSO
AU - Pan, Zhaoye
AU - Wang, Xiaolu
AU - Liu, Huikang
AU - Zhang, Jun
N1 - Publisher Copyright:
© 2024 IEEE.
PY - 2024
Y1 - 2024
N2 - Time-varying graphical LASSO (TVGL) aims to infer a sequence of graphs from time series data and has been widely used in many statistical inference problems. The existing algorithms usually suffer from high computational cost when solving large-scale TVGL problems. In this paper, we develop an efficient and scalable hierarchical block coordinate descent (HBCD) method for solving TVGL with smooth temporal difference prior. The proposed HBCD method contains both outer-loop and inner-loop BCD iterations. The outer loops seperate the original TVGL problem into a sequence of subproblems, which are variants of the static graphical LASSO problems. Then, we propose an efficient BCD method to solve the inner-loop subproblems. We provide theoretical analysis that indicates the linear convergence of our proposed method. Furthermore, numerical experiments on both synthetic and real datasets show that our method significantly outperforms the state-of-the-art algorithm in terms of both the required iterations and CPU time to reach the target precision.
AB - Time-varying graphical LASSO (TVGL) aims to infer a sequence of graphs from time series data and has been widely used in many statistical inference problems. The existing algorithms usually suffer from high computational cost when solving large-scale TVGL problems. In this paper, we develop an efficient and scalable hierarchical block coordinate descent (HBCD) method for solving TVGL with smooth temporal difference prior. The proposed HBCD method contains both outer-loop and inner-loop BCD iterations. The outer loops seperate the original TVGL problem into a sequence of subproblems, which are variants of the static graphical LASSO problems. Then, we propose an efficient BCD method to solve the inner-loop subproblems. We provide theoretical analysis that indicates the linear convergence of our proposed method. Furthermore, numerical experiments on both synthetic and real datasets show that our method significantly outperforms the state-of-the-art algorithm in terms of both the required iterations and CPU time to reach the target precision.
KW - Hierarchical Block Coordinate Descent
KW - Laplacian Penalty
KW - Time-Varying Graphical LASSO
UR - https://www.scopus.com/pages/publications/85195378043
U2 - 10.1109/ICASSP48485.2024.10446070
DO - 10.1109/ICASSP48485.2024.10446070
M3 - 会议稿件
AN - SCOPUS:85195378043
T3 - ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
SP - 9836
EP - 9840
BT - 2024 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2024 - Proceedings
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2024 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2024
Y2 - 14 April 2024 through 19 April 2024
ER -