On the computational power of neural nets

  title={On the computational power of neural nets},
  author={H. Siegelmann and Eduardo Sontag},
  booktitle={COLT '92},
  • H. Siegelmann, Eduardo Sontag
  • Published in COLT '92 1992
  • Computer Science
  • This paper deals with finite networks which consist of interconnections of synchronously evolving processors. Each processor updates its state by applying a “sigmoidal” scalar nonlinearity to a linear combination of the previous states of all units. We prove that one may simulate all Turing Machines by rational nets. In particular, one can do this in linear time, and there is a net made up of about 1,000 processors which computes a universal partial-recursive function. Products (high order nets… CONTINUE READING
    Analog Computation via Neural Networks
    • 362
    • PDF
    Lower Bounds for the Computational Power of Networks of Spiking Neurons
    • W. Maass
    • Mathematics, Computer Science
    • 1996
    • 225
    • PDF
    Extraction of rules from discrete-time recurrent neural networks
    • 194
    • PDF
    On Alan Turing's anticipation of connectionism
    • 59
    • PDF
    Optimal Simulation of Automata by Neural Nets
    • 50
    Computational complexity of neural networks: a survey
    • 71
    On the power of sigmoid neural networks
    • 26
    • PDF