Non-negative Matrix Factorization under Heavy Noise

被引:0
|
作者
Bhattacharya, Chiranjib [1 ]
Goyal, Navin [2 ]
Kannan, Ravindran [2 ]
Pani, Jagdeep [1 ]
机构
[1] Indian Inst Sci, Dept Comp Sci & Automat, Bangalore, Karnataka, India
[2] Microsoft Res India, Bangalore, Karnataka, India
关键词
ALGORITHM;
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
The Noisy Non-negative Matrix factorization (NMF) is: given a data matrix A (d x n), find non-negative matrices B;C (d x k, k x n respy.) so that A = BC + N, where N is a noise matrix. Existing polynomial time algorithms with proven error guarantees require each column N-x;j to have l(1) norm much smaller than parallel to(BC)(x;j)parallel to(1), which could be very restrictive. In important applications of NMF such as Topic Modeling as well as theoretical noise models (eg. Gaussian with high sigma), almost every column of N-xj violates this condition. We introduce the heavy noise model which only requires the average noise over large subsets of columns to be small. We initiate a study of Noisy NMF under the heavy noise model. We show that our noise model subsumes noise models of theoretical and practical interest (for eg. Gaussian noise of maximum possible sigma). We then devise an algorithm TSVDNMF which under certain assumptions on B;C, solves the problem under heavy noise. Our error guarantees match those of previous algorithms. Our running time of O ((n + d)(2)k) is substantially better than the O(n(3)d) for the previous best. Our assumption on B is weaker than the "Separability" assumption made by all previous results. We provide empirical justification for our assumptions on C. We also provide the first proof of identifiability (uniqueness of B) for noisy NMF which is not based on separability and does not use hard to check geometric conditions. Our algorithm outperforms earlier polynomial time algorithms both in time and error, particularly in the presence of high noise.
引用
收藏
页数:9
相关论文
共 50 条
  • [1] Non-negative Matrix Factorization for Images with Laplacian Noise
    Lam, Edmund Y.
    2008 IEEE ASIA PACIFIC CONFERENCE ON CIRCUITS AND SYSTEMS (APCCAS 2008), VOLS 1-4, 2008, : 798 - 801
  • [2] Dropout non-negative matrix factorization
    Zhicheng He
    Jie Liu
    Caihua Liu
    Yuan Wang
    Airu Yin
    Yalou Huang
    Knowledge and Information Systems, 2019, 60 : 781 - 806
  • [3] Non-negative matrix factorization on kernels
    Zhang, Daoqiang
    Zhou, Zhi-Hua
    Chen, Songcan
    PRICAI 2006: TRENDS IN ARTIFICIAL INTELLIGENCE, PROCEEDINGS, 2006, 4099 : 404 - 412
  • [4] Non-negative Matrix Factorization: A Survey
    Gan, Jiangzhang
    Liu, Tong
    Li, Li
    Zhang, Jilian
    COMPUTER JOURNAL, 2021, 64 (07): : 1080 - 1092
  • [5] Collaborative Non-negative Matrix Factorization
    Benlamine, Kaoutar
    Grozavu, Nistor
    Bennani, Younes
    Matei, Basarab
    ARTIFICIAL NEURAL NETWORKS AND MACHINE LEARNING - ICANN 2019: TEXT AND TIME SERIES, PT IV, 2019, 11730 : 655 - 666
  • [6] INFINITE NON-NEGATIVE MATRIX FACTORIZATION
    Schmidt, Mikkel N.
    Morup, Morten
    18TH EUROPEAN SIGNAL PROCESSING CONFERENCE (EUSIPCO-2010), 2010, : 905 - 909
  • [7] Non-negative Matrix Factorization for EEG
    Jahan, Ibrahim Salem
    Snasel, Vaclav
    2013 INTERNATIONAL CONFERENCE ON TECHNOLOGICAL ADVANCES IN ELECTRICAL, ELECTRONICS AND COMPUTER ENGINEERING (TAEECE), 2013, : 183 - 187
  • [8] Algorithms for non-negative matrix factorization
    Lee, DD
    Seung, HS
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 13, 2001, 13 : 556 - 562
  • [9] Non-negative matrix factorization with α-divergence
    Cichocki, Andrzej
    Lee, Hyekyoung
    Kim, Yong-Deok
    Choi, Seungjin
    PATTERN RECOGNITION LETTERS, 2008, 29 (09) : 1433 - 1440
  • [10] Dropout non-negative matrix factorization
    He, Zhicheng
    Liu, Jie
    Liu, Caihua
    Wang, Yuan
    Yin, Airu
    Huang, Yalou
    KNOWLEDGE AND INFORMATION SYSTEMS, 2019, 60 (02) : 781 - 806