Latent Variable Autoencoder

doi:10.1109/ACCESS.2019.2910152

	Latent Variable Autoencoder
	Wenjuan Han 1; Ge Wang 1; Kewei Tu2
	2019
发表期刊	IEEE ACCESS
ISSN	2169-3536
卷号	7 期号:99 页码:48514-48523
发表状态	已发表
DOI	10.1109/ACCESS.2019.2910152
摘要	Learning to discover hidden variables from unlabeled data is an important task. Traditional generative methods model the generation process of the observed variables as well as the hidden variables. However, tractable inference and learning on these models requires strong conditional independence assumptions being made among observed and hidden variables. To tackle this limitation, we propose an autoencoder framework. The encoder produces an intermediate representation from the observed variables, and the decoder is a generative latent variable model conditioned on the intermediate representation that tries to generate the hidden variables as well as to reconstruct the observed variables. We introduce three variant models of our framework with either a deterministic or a stochastic encoding process. To optimize our model, we propose an algorithm similar to the classic expectation-maximization (EM) algorithm that supports online learning for large-scale datasets. The flexibility of our framework allows us to apply it to various scenarios where the explicit inference of hidden variables is desired. We discuss the applications of our framework to the perceptual grouping task and the part-of-speech (POS) induction task. Our experiments on the two tasks demonstrate that our framework can achieve better performance than vanilla latent variable generative models.
关键词	Unsupervised learning autoencoder neural networks
URL	查看原文
收录类别	SCI ; SCIE ; EI
语种	英语
WOS研究方向	Computer Science ; Engineering ; Telecommunications
WOS类目	Computer Science, Information Systems ; Engineering, Electrical & Electronic ; Telecommunications
WOS记录号	WOS:000467526700001
出版者	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
EI入藏号	20191906880422
EI主题词	Large dataset ; Learning systems ; Maximum principle ; Neural networks ; Signal encoding ; Stochastic systems ; Unsupervised learning
EI分类号	Information Theory and Signal Processing:716.1 ; Probability Theory:922.1 ; Systems Science:961
原始文献类型	Article
来源库	IEEE
引用统计
文献类型	期刊论文
条目标识符	https://kms.shanghaitech.edu.cn/handle/2MSLDSTB/31168
专题	信息科学与技术学院信息科学与技术学院_PI研究组_屠可伟组
作者单位	1.University of Chinese Academy of Sciences, Beijing, China 2.School of Information Science and Technology, ShanghaiTech University, Shanghai, China
推荐引用方式 GB/T 7714	Wenjuan Han,Ge Wang,Kewei Tu. Latent Variable Autoencoder[J]. IEEE ACCESS,2019,7(99):48514-48523.
APA	Wenjuan Han,Ge Wang,&Kewei Tu.(2019).Latent Variable Autoencoder.IEEE ACCESS,7(99),48514-48523.
MLA	Wenjuan Han,et al."Latent Variable Autoencoder".IEEE ACCESS 7.99(2019):48514-48523.