Robust Bayesian matrix decomposition with mixture of Gaussian noise

doi:10.1016/j.neucom.2021.04.004

CSpace > 应用数学研究所

	Robust Bayesian matrix decomposition with mixture of Gaussian noise
	Wang, Haohui 1; Zhang, Chihao 2,3; Zhang, Shihua2,3
	2021-08-18
发表期刊	NEUROCOMPUTING
ISSN	0925-2312
卷号	449 页码:108-116
摘要	Matrix decomposition is a popular and fundamental approach in machine learning. The classical matrix decomposition methods with Frobenius norm loss is only optimal for Gaussian noise and thus suffer from the sensitivity to outliers and non-Gaussian noise. To address these limitations, the proposed methods can be divided into two categories. One type of approach is to replace the Frobenius norm loss with robust loss functions. The other type of approach is to impose the Bayesian priors to reduce the risk of overfitting. This paper combines these two approaches. Specifically, we model the noise by a mixture of Gaussian distribution, enabling the model to approximate a wide range of noise distributions. Meanwhile, we put a Laplace prior on the basis matrix to enforce the sparsity and a Dirichlet prior on the coefficient matrix to improve the interpretability. Extensive experiments in synthetic data and real-world data demonstrate that this method outperforms several competing ones. Ablation studies show that this method benefits from both the Bayesian priors and the Mixture of Gaussian noise loss, which confirms the necessity of combining the two schemes. (c) 2021 Elsevier B.V. All rights reserved.
关键词	Bayesian method Matrix decomposition Maximum a posterior Mixture of Gaussians
DOI	10.1016/j.neucom.2021.04.004
收录类别	SCI
语种	英语
资助项目	National Key Research and Development Program of China[2019YFA0709501] ; National Natural Science Foundation of China[11661141019] ; National Natural Science Foundation of China[61621003] ; National Ten Thousand Talent Program for Young Topnotch Talents ; CAS Frontier Science Research Key Project for Top Young Scientist[QYZDBSSWSYS008]
WOS研究方向	Computer Science
WOS类目	Computer Science, Artificial Intelligence
WOS记录号	WOS:000652818400010
出版者	ELSEVIER
引用统计
文献类型	期刊论文
条目标识符	http://ir.amss.ac.cn/handle/2S8OKBNM/58708
专题	应用数学研究所
通讯作者	Zhang, Shihua
作者单位	1.Zhejiang Univ, Sch Math Sci, Hangzhou 310027, Peoples R China 2.Chinese Acad Sci, NCMIS, CEMS, RCSDS,Acad Math & Syst Sci, Beijing 100190, Peoples R China 3.Univ Chinese Acad Sci, Sch Math Sci, Beijing 100049, Peoples R China
推荐引用方式 GB/T 7714	Wang, Haohui,Zhang, Chihao,Zhang, Shihua. Robust Bayesian matrix decomposition with mixture of Gaussian noise[J]. NEUROCOMPUTING,2021,449:108-116.
APA	Wang, Haohui,Zhang, Chihao,&Zhang, Shihua.(2021).Robust Bayesian matrix decomposition with mixture of Gaussian noise.NEUROCOMPUTING,449,108-116.
MLA	Wang, Haohui,et al."Robust Bayesian matrix decomposition with mixture of Gaussian noise".NEUROCOMPUTING 449(2021):108-116.