Skip to main content
SpringerLink
Log in

The “bottleneck” behaviours in linear feedforward neural network classifiers and their breakthrough

  • Correspondence
  • Published:
Journal of Computer Science and Technology Aims and scope Submit manuscript

Abstract

The classification mechanisms of linear feedforward neural network classifiers (FNNC), whose hidden layer performs the Fisher linear transformation of the input patterns, under the supervision of outer-supervised signals are investigated. The “bottleneck” behaviours in linear FNNCs are observed and analyzed. In addition, the structure stabilities of the linear FNNCs are also discussed. It is pointed out that the key point to break through the “bottleneck” behaviours for linear FNNCs is to change linear hidden neurons into nonlinear hidden ones. Finally, the experimental results, taking the parity 3 problem as example, are given.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Similar content being viewed by others

References

  1. Gallinari P, Thiria S, Soulie F F. Multilayer perceptrons and data analysis. In2nd IJCNN, 1988, 1: 391–399.

  2. Gallinari P, Thiria S, Badran F, Soulie F F. On the relations between discriminant analysis and multilayer perceptrons.Neural Networks, 1991, 4(2): 349–360.

    Article  Google Scholar 

  3. Lu X X, Yuan K P, Huang D S. A study of backpropagation learning algorithm of multilayer perceptron networks based on recursive least squares. InProc. the Sixth Japan-China International Conference on Cmputer Applications, Sept. 16–18, Sapporo, Japan, 1994, pp. 169–172 (J-4).

  4. Hush D R, Salas J M. Improving the learning rate of backpropagation with gradient reuse algorithm.IJCNN (Washington D C), 1988, I: 441–447.

    Google Scholar 

  5. Huang D S. Linear and nonlinear feedforward neural network classifiers: a comprehensive understanding.Journal of Intelligent Systems, 1998, 8(3–4).

  6. Duda R O. Pattern Classification and Scene Analysis. John Wiley & Sons, Inc., 1973.

  7. Huang D S. The local minima free condition of feedforward neural networks for outer-supervised learning.IEEE Trans. Systems, Man and Cybernetics, 1988, 28B(3): 477–480.

    Google Scholar 

  8. Pao Y H. Adaptive Pattern Recognition and Neural Networks. Addison-Wesley Publishing Company Inc., 1989.

Download references

Author information

Authors and Affiliations

  1. Beijing Institute of System Engineering, P.O. Box 9702-19, 100101, Beijing, P.R. China

    Huang Deshuang

Authors
  1. Huang Deshuang
    View author publications

    You can also search for this author in PubMed Google Scholar

Additional information

This work was supported by the National Natural Science Foundation of China under the grant No. 69705001.

HUANG Deshuang was born in 1964. He received the M.S. and Ph.D. degrees in electrical engineering from the National University of Defense Technology, Changsha and Xidian University, Xian, in 1989 and 1993, respectively. He did research work as a Postdoctor research fellow at Beijing Institute of Technology, Beijing and National Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences, respectively, from July 1993 to August 1997. He has published over 80 papers and one book. His research interests include signal processing, pattern recognition, neural networks and fuzzy logic.

Rights and permissions

Reprints and permissions

About this article

Cite this article

Huang, D. The “bottleneck” behaviours in linear feedforward neural network classifiers and their breakthrough. J. Comput. Sci. & Technol. 14, 34–43 (1999). https://doi.org/10.1007/BF02952485

Download citation

  • Received:

  • Revised:

  • Issue Date:

  • DOI: https://doi.org/10.1007/BF02952485

Keywords

Search

Navigation