Genetic Programming Techniques that Evolve Recurrent Neural Networks Architectures for Signal Processing

Created by W.Langdon from gp-bibliography.bib Revision:1.2031

@InProceedings{esparcia:1996:GPerNNasp,
  author =       "Anna I. Esparcia-Alcazar and Kenneth C. Sharman",
  title =        "Genetic Programming Techniques that Evolve Recurrent
                 Neural Networks Architectures for Signal Processing",
  booktitle =    "IEEE Workshop on Neural Networks for Signal
                 Processing",
  year =         "1996",
  month =        "4-6 " # sep,
  pages =        "139--148",
  address =      "Seiko, Kyoto, Japan",
  publisher =    "IEEE",
  keywords =     "genetic algorithms, genetic programming, adaptive
                 filtering, arbitrary transfer functions, design
                 constraints, genetic programming techniques, neuronal
                 transfer functions, online training algorithm,
                 recurrent neural network architectures, signal
                 processing, simulated annealing, adaptive filters,
                 geometric programming, neural net architecture,
                 recurrent neural nets, signal processing, simulated
                 annealing, transfer functions",
  doi =          "doi:10.1109/NNSP.1996.548344",
  size =         "10 pages",
  abstract =     "We propose a novel design paradigm for recurrent
                 neural networks. This employs a two-stage genetic
                 programming/simulated annealing hybrid algorithm to
                 produce a neural network which satisfies a set of
                 design constraints. The genetic programming part of the
                 algorithm is used to evolve the general topology of the
                 network, along with specifications for the neuronal
                 transfer functions, while the simulated annealing
                 component of the algorithm adapts the network's
                 connection weights in response to a set of training
                 data. Our approach offers important advantages over
                 existing methods for automated network design. Firstly,
                 it allows us to develop recurrent network connections;
                 secondly, we are able to employ neurones with arbitrary
                 transfer functions, and thirdly, the approach yields an
                 efficient and easy to implement on-line training
                 algorithm. The procedures involved in using the GP/SA
                 hybrid algorithm are illustrated by using it to design
                 a neural network for adaptive filtering in a signal
                 processing application",
}

Genetic Programming entries for Anna Esparcia-Alcazar Kenneth C Sharman