Classification of all Stationary Points on a Neural Network Error Surface

@inproceedings{SprinkhuizenKuyper1994ClassificationOA,
  title={Classification of all Stationary Points on a Neural Network Error Surface},
  author={Ida G. Sprinkhuizen-Kuyper and Egbert J. W. Boers},
  year={1994}
}
The artificial neural network with one hidden unit and the input nodes connected to the output node is considered. It is proven that the error surface of this network for the patterns of the XOR problem has minimum values with zero error and that all other stationary points of the error surface are saddle points. Also, the volume of the regions in weight space with saddle points is zero, hence training this network, using e.g. backpropagation with momentum, on the four patterns of the XOR… CONTINUE READING
4 Citations
14 References
Similar Papers

References

Publications referenced by this paper.
Showing 1-10 of 14 references

Perantonis; “Complete solution of the local minima in the XOR problem”,Network

  • S.J.P.J.G. Lisboa
  • 1991
Highly Influential
3 Excerpts

Biological metaphors in designing modular artificial neural networks

  • E.J.W. Boers, H. Kuiper, B.L.M. Happel, I. G. Sprinkhuizen-Kuyper
  • Proceedings of the International Conference on…
  • 1993
1 Excerpt

Sprinkhuizen-Kuyper; “Designing Modular Artificial Neural Networks

  • E.J.W. Boers, H. Kuiper, I.G.B.L.M. Happel
  • H.A. Wijshoff (ed.); Proceedings of Computing…
  • 1993
1 Excerpt

Taylor; “Avoiding Local Minima by Progressive Range Expansion

  • D. Gorse, J.G.A. Shepherd
  • Proceedings of the International Conference on…
  • 1993
1 Excerpt

Supervised Learning: A Theoretical Framework

  • S. A. Solla
  • Proc. Vol. XII , Addison-Wesley,
  • 1992
1 Excerpt

Complete solution of the local minima in the XOR problem ”

  • S. J. Perantonis
  • 1991

Palmer; Introduction to the Theory of Neural Computation

  • J. Herz, R.G.A. Krogh
  • 1991

Denker: “Exhaustive Learning

  • D. B. Schwartz, V. K. Samalam, J.S.S.A. Solla
  • Neural Computation
  • 1990
1 Excerpt

Similar Papers

Loading similar papers…