作者其他论文
文献详情
Regularized Non-negative Matrix Factorization Using Alternating Direction Method of Multipliers and Its Application to Source Separation
文献类型:会议
作者:Zhang, Shaofei[1]  Huang, Dongyan[2]  Xie, Lei[3]  Chng, Eng Siong[4]  Li, Haizhou[5]  Dong, Minghui[6]  
机构:[1]Northwestern Polytech Univ, Sch Comp Sci, Xian, Peoples R China.;
[2]ASTAR, Inst Infocomm Res, Singapore, Singapore.;
[3]Northwestern Polytech Univ, Sch Comp Sci, Xian, Peoples R China.;
[4]Nanyang Technol Univ, Sch Comp Engn, Singapore, Singapore.;
[5]ASTAR, Inst Infocomm Res, Singapore, Singapore.;Nanyang Technol Univ, Sch Comp Engn, Singapore, Singapore.;
[6]ASTAR, Inst Infocomm Res, Singapore, Singapore.;
年:2015
通讯作者:Zhang, SF (reprint author), Northwestern Polytech Univ, Sch Comp Sci, Xian, Peoples R China.
会议名称:16th Annual Conference of the International-Speech-Communication-Association (INTERSPEECH 2015)
会议论文集:16TH ANNUAL CONFERENCE OF THE INTERNATIONAL SPEECH
页码范围:1498-1502
会议地点:Dresden, GERMANY
会议开始日期:2015-01-01
收录情况:CPCI-S(WOS:000380581600313)  
所属部门:计算机学院
人气指数:641
浏览次数:631
语言:外文
关键词:Regularized non-negative matrix factorization; Alternating direction method of multipliers; Beta-divergence; Source separation
摘要:
Non-negative matrix factorization (NMF) aims at finding non negative representations of nonnegative data. Among different NMF algorithms, alternating direction method of multipliers (ADMM) is a popular one with superior performance. However, we find that ADMM shows instability and inferior performance on real-world data like speech signals. In this paper, to solve this problem, we develop a class of advanced regularized ADMM algorithms for NMF. Efficient and robust learning rules are achieved by ...More
0
评论(0 条评论)
登录