Non-negative matrix factorization using stable alternating direction method of multipliers for source separation

Shaofei Zhang, Dongyan Huang, Lei Xie, Eng Siong Chng, Haizhou Li, Minghui Dong

科研成果: 书/报告/会议事项章节会议稿件同行评审

7 引用 (Scopus)

摘要

Nonnegative matrix factorization (NMF) is a popular method for source separation. In this paper, an alternating direction method of multipliers (ADMM) for NMF is studied, which deals with the NMF problem using the cost function of beta-divergence. Our study shows that this algorithm outperforms state-of-the-art algorithms on synthetic data sets, but it presents unstable behavior and low accuracy on real data sets. Therefore, we propose two different stable ADMM algorithms for NMF to solve this problem. They differ slightly in the multiplicative factor utilized in the update rules. One algorithm is to adapt the step size to guarantee the convergence while the other minimizes the beta-divergence with a pivot element weighting iterative method (PEWI). Experimental results demonstrate that the proposed algorithms are more stable and accurate. Particularly, PEWI based ADMM shows superior performance in the source separation task.

源语言英语
主期刊名2015 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2015
出版商Institute of Electrical and Electronics Engineers Inc.
222-228
页数7
ISBN(电子版)9789881476807
DOI
出版状态已出版 - 19 2月 2016
活动2015 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2015 - Hong Kong, 香港
期限: 16 12月 201519 12月 2015

出版系列

姓名2015 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2015

会议

会议2015 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2015
国家/地区香港
Hong Kong
时期16/12/1519/12/15

指纹

探究 'Non-negative matrix factorization using stable alternating direction method of multipliers for source separation' 的科研主题。它们共同构成独一无二的指纹。

引用此