Gradient Neural Network with Nonlinear Activation for Computing Inner Inverses and the Drazin Inverse

被引：0
|
作者：
Predrag S. Stanimirović
Marko D. Petković
Dimitrios Gerontitis
机构：
[1] University of Niš,Faculty of Sciences and Mathematics
[2] Aristoteleion Panepistimion,undefined
来源：
Neural Processing Letters | 2018年 / 48卷
关键词：
Recurrent neural network; Moore–Penrose inverse; Drazin inverse; Dynamic equation; Activation function; 68T05; 15A09; 65F20;
D O I：
暂无
中图分类号：
学科分类号：
摘要：
Two gradient-based recurrent neural networks (GNNs) for solving two matrix equations are presented and investigated. These GNNs can be used for generating various inner inverses, including the Moore–Penrose, and in the computation of the Drazin inverse. Convergence properties of defined GNNs are considered. Certain conditions which impose convergence towards the pseudoinverse, and the Drazin inverse are exactly specified. The influence of nonlinear activation functions on the convergence performance of defined GNN models is investigated. Computer simulation experience further confirms the theoretical results.
引用
收藏
页码：109 / 133
页数：24
相关论文
共 50 条
  • [1] Gradient Neural Network with Nonlinear Activation for Computing Inner Inverses and the Drazin Inverse
    Stanimirovic, Predrag S.
    Petkovic, Marko D.
    Gerontitis, Dimitrios
    NEURAL PROCESSING LETTERS, 2018, 48 (01) : 109 - 133
  • [2] Recurrent Neural Network for Computing the Drazin Inverse
    Stanimirovic, Predrag S.
    Zivkovic, Ivan S.
    Wei, Yimin
    IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS, 2015, 26 (11) : 2830 - 2843
  • [3] Nonlinearly Activated Recurrent Neural Network for Computing the Drazin Inverse
    Xue-Zhong Wang
    Haifeng Ma
    Predrag S. Stanimirović
    Neural Processing Letters, 2017, 46 : 195 - 217
  • [4] Nonlinearly Activated Recurrent Neural Network for Computing the Drazin Inverse
    Wang, Xue-Zhong
    Ma, Haifeng
    Stanimirovic, Predrag S.
    NEURAL PROCESSING LETTERS, 2017, 46 (01) : 195 - 217
  • [5] Recurrent neural network for computing the W-weighted Drazin inverse
    Wang, Xue-Zhong
    Ma, Haifeng
    Stanimirovic, Predrag S.
    APPLIED MATHEMATICS AND COMPUTATION, 2017, 300 : 1 - 20
  • [6] Gradient methods for computing the Drazin-inverse solution
    Miljkovic, Sladjana
    Miladinovic, Marko
    Stanimirovic, Predrag S.
    Wei, Yimin
    JOURNAL OF COMPUTATIONAL AND APPLIED MATHEMATICS, 2013, 253 : 255 - 263
  • [7] Recurrent Neural Network Approach Based on the Integral Representation of the Drazin Inverse
    Stanimirovic, Predrag S.
    Zivkovic, Ivan S.
    Wei, Yimin
    NEURAL COMPUTATION, 2015, 27 (10) : 2107 - 2131
  • [8] Complex Neural Network Models for Time-Varying Drazin Inverse
    Wang, Xue-Zhong
    Wei, Yimin
    Stanimirovic, Predrag S.
    NEURAL COMPUTATION, 2016, 28 (12) : 2790 - 2824
  • [9] Recurrent Neural Network for Computing Outer Inverse
    Zivkovic, Ivan S.
    Stanimirovic, Predrag S.
    Wei, Yimin
    NEURAL COMPUTATION, 2016, 28 (05) : 970 - 998
  • [10] Neural network approach to computing outer inverses based on the full rank representation
    Stanimirovic, Predrag S.
    Zivkovic, Ivan S.
    Wei, Yimin
    LINEAR ALGEBRA AND ITS APPLICATIONS, 2016, 501 : 344 - 362
12345