Controlling sparseness in non-negative tensor factorization

被引：0
|
作者：
Heiler, M [1 ]
Schnörr, C [1 ]
机构：
[1] Univ Mannheim, Dept Math & Comp Sci, Comp Vis Graph & Pattern Recognit Grp, D-68131 Mannheim, Germany
来源：
COMPUTER VISION - ECCV 2006 , PT 1, PROCEEDINGS | 2006年 / 3951卷
关键词：
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：
Non-negative tenser factorization (NTF) has recently been proposed as sparse and efficient image representation (Welling and Weber, Patt. Rec. Let., 2001). Until now, sparsity of the tensor factorization has been empirically observed in many cases, but there was no systematic way to control it. In this work, we show that a sparsity measure recently proposed for non-negative matrix factorization (Hoyer, J. Mach. Learn. Res., 2004) applies to NTF and allows precise control over sparseness of the resulting factorization. We devise an algorithm based on sequential conic programming and show improved performance over classical NTF codes on artificial and on real-world data sets.
引用
收藏
页码：56 / 67
页数：12
相关论文
共 50 条
  • [21] Overlapping Community Detection Using Non-Negative Matrix Factorization With Orthogonal and Sparseness Constraints
    Chen, Naiyue
    Liu, Yun
    Chao, Han-Chieh
    IEEE ACCESS, 2018, 6 : 21266 - 21274
  • [22] Non-negative Einstein tensor factorization for unmixing hyperspectral images
    El Hachimi, Anas
    Jbilou, Khalide
    Ratnani, Ahmed
    NUMERICAL ALGORITHMS, 2025,
  • [23] Non-Negative Tensor Factorization Applied to Music Genre Classification
    Benetos, Emmanouil
    Kotropoulos, Constantine
    IEEE TRANSACTIONS ON AUDIO SPEECH AND LANGUAGE PROCESSING, 2010, 18 (08): : 1955 - 1967
  • [24] QUASI-ORTHOGONALIZATION FOR ALTERNATING NON-NEGATIVE TENSOR FACTORIZATION
    Grasedyck, Lars
    Klever, Maren
    Kraemer, Sebastian
    ELECTRONIC TRANSACTIONS ON NUMERICAL ANALYSIS, 2024, 62 : 22 - 57
  • [25] A sufficient condition for the unique solution of non-negative tensor factorization
    Sumi, Toshio
    Sakata, Toshio
    INDEPENDENT COMPONENT ANALYSIS AND SIGNAL SEPARATION, PROCEEDINGS, 2007, 4666 : 113 - +
  • [26] NEIGHBORHOOD PRESERVING NON-NEGATIVE TENSOR FACTORIZATION FOR IMAGE REPRESENTATION
    Wang, Yu-Xiong
    Gui, Liang-Yan
    Zhang, Yu-Jin
    2012 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), 2012, : 3389 - 3392
  • [27] VARIATIONAL BAYESIAN PARTIALLY OBSERVED NON-NEGATIVE TENSOR FACTORIZATION
    Hinrich, Jesper L.
    Nielsen, Soren F. V.
    Madsen, Kristoffer H.
    Morup, Morten
    2018 IEEE 28TH INTERNATIONAL WORKSHOP ON MACHINE LEARNING FOR SIGNAL PROCESSING (MLSP), 2018,
  • [28] Molecular cancer class discovery using non-negative matrix factorization with sparseness constraint
    Kong, Xiangzhen
    Zheng, Chunhou
    Wu, Yuqiang
    Shang, Li
    ADVANCED INTELLIGENT COMPUTING THEORIES AND APPLICATIONS: WITH ASPECTS OF THEORETICAL AND METHODOLOGICAL ISSUES, 2007, 4681 : 792 - +
  • [29] Topic Diffusion Discovery based on Sparseness-constrained Non-negative Matrix Factorization
    Kang, Yihuang
    Lin, Keng-Pei
    Cheng, I-Ling
    2018 IEEE INTERNATIONAL CONFERENCE ON INFORMATION REUSE AND INTEGRATION (IRI), 2018, : 94 - 101
  • [30] Percussive/harmonic sound separation by non-negative matrix factorization with smoothness/sparseness constraints
    Jesus Canadas-Quesada, Francisco
    Vera-Candeas, Pedro
    Ruiz-Reyes, Nicolas
    Carabias-Orti, Julio
    Cabanas-Molero, Pablo
    EURASIP JOURNAL ON AUDIO SPEECH AND MUSIC PROCESSING, 2014, : 1 - 17
12345