Statistical Physics and Practical Training of Soft – Committee Machines

  title={Statistical Physics and Practical Training of Soft – Committee Machines},
  author={Martin Ahr and Michael Biehl and Robert Urbanczik},
Equilibrium states of large layered neural networks with differentiable activation function and a single, linear output unit are investigated using the replica formalism. The quenched free energy of a student network with a very large number of hidden units learning a rule of perfectly matching complexity is calculated analytically. The system undergoes a first order phase transition from unspecialized to specialized student configurations at a critical size of the training set. Computer… CONTINUE READING
4 Citations
5 References
Similar Papers


Publications referenced by this paper.
Showing 1-5 of 5 references

Neural Networks for Pattern Recognition, (Clarendon

  • C. Bishop
  • 1995
Highly Influential
4 Excerpts

A 28

  • B. Schottky, J. Phys
  • 4515
  • 1995
1 Excerpt


  • R. Urbanczik
  • Phys A 28, 7097
  • 1995
5 Excerpts

Rumelhart (eds.), Backpropagation: Theory, Architecture, and Applications (Lawrence Erlbaum

  • D.E.Y. Chauvin
  • 1995
2 Excerpts

Introduction to the Theory of Neural Com- putation, (Addison{Wesley

  • J. A. Hertz, A. Krogh R.G. Palmer
  • 1991
3 Excerpts

Similar Papers

Loading similar papers…