Scalable Inference of Sparsely-changing Gaussian Markov Random Fields

被引:0
作者
Fattahi, Salar [1 ]
Gomez, Andres [2 ]
机构
[1] Univ Michigan, Dept Ind & Operat Engn, Ann Arbor, MI 48109 USA
[2] Univ Southern Calif, Dept Ind & Syst Engn, Los Angeles, CA 90089 USA
来源
ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 34 (NEURIPS 2021) | 2021年 / 34卷
基金
美国国家科学基金会;
关键词
DECAY BOUNDS; COVARIANCE; ALGORITHM; MATRICES; INVERSE; RATES;
D O I
暂无
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
We study the problem of inferring time-varying Gaussian Markov random fields, where the underlying graphical model is both sparse and changes sparsely over time. Most of the existing methods for the inference of time-varying Markov random fields (MRFs) rely on the regularized maximum likelihood estimation (MLE), that typically suffer from weak statistical guarantees and high computational time. Instead, we introduce a new class of constrained optimization problems for the inference of sparsely-changing Gaussian MRFs (GMRFs). The proposed optimization problem is formulated based on the exact l(0) regularization, and can be solved in near-linear time and memory. Moreover, we show that the proposed estimator enjoys a provably small estimation error. We derive sharp statistical guarantees in the high-dimensional regime, showing that such problems can be learned with as few as one sample per time period. Our proposed method is extremely efficient in practice: it can accurately estimate sparsely-changing GMRFs with more than 500 million variables in less than one hour.
引用
收藏
页数:13
相关论文
共 50 条
[21]   GLOBAL MINIMIZATION OF MARKOV RANDOM FIELDS WITH APPLICATIONS TO OPTICAL FLOW [J].
Goldstein, Tom ;
Bresson, Xavier ;
Osher, Stan .
INVERSE PROBLEMS AND IMAGING, 2012, 6 (04) :623-644
[22]   Diverse M-Best Solutions in Markov Random Fields [J].
Batra, Dhruv ;
Yadollahpour, Payman ;
Guzman-Rivera, Abner ;
Shakhnarovich, Gregory .
COMPUTER VISION - ECCV 2012, PT V, 2012, 7576 :1-16
[23]   Gaussian Conditional Random Fields for Aggregation of Operational Aerosol Retrievals [J].
Djuric, Nemanja ;
Radosavljevic, Vladan ;
Obradovic, Zoran ;
Vucetic, Slobodan .
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS, 2015, 12 (04) :761-765
[24]   Universal Inference Meets Random Projections: A Scalable Test for Log-Concavity [J].
Dunn, Robin ;
Gangrade, Aditya ;
Wasserman, Larry ;
Ramdas, Aaditya .
JOURNAL OF COMPUTATIONAL AND GRAPHICAL STATISTICS, 2025, 34 (01) :267-279
[25]   Generalized linear mixed models with Gaussian mixture random effects: Inference and application [J].
Pan, Lanfeng ;
Li, Yehua ;
He, Kevin ;
Li, Yanming ;
Li, Yi .
JOURNAL OF MULTIVARIATE ANALYSIS, 2020, 175
[26]   Global optimization for first order Markov Random Fields with submodular priors [J].
Darbon, Jerome .
DISCRETE APPLIED MATHEMATICS, 2009, 157 (16) :3412-3423
[27]   Integrated Denoising and Unwrapping of InSAR Phase Based on Markov Random Fields [J].
Chen, Runpu ;
Yu, Weidong ;
Wang, Robert ;
Liu, Gang ;
Shao, Yunfeng .
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, 2013, 51 (08) :4473-4485
[28]   Similarity/dissimilarity analysis of protein structures based on Markov random fields [J].
Wu, Jiaqi ;
Zhou, Tao ;
Tao, Jin ;
Hai, Yabing ;
Ye, Fei ;
Liu, Xiaoqing ;
Dai, Qi .
COMPUTATIONAL BIOLOGY AND CHEMISTRY, 2018, 75 :45-53
[29]   Convergence of Markovian stochastic approximation for Markov random fields with hidden variables [J].
Qi, Anna ;
Yang, Lihua ;
Huang, Chao .
STOCHASTICS AND DYNAMICS, 2020, 20 (05)
[30]   Disease gene identification by using graph kernels and Markov random fields [J].
Chen BoLin ;
Li Min ;
Wang JianXin ;
Wu FangXiang .
SCIENCE CHINA-LIFE SCIENCES, 2014, 57 (11) :1054-1063