Approximation capabilities of adaptive spline neural networks

@article{Vecci1997ApproximationCO,
  title={Approximation capabilities of adaptive spline neural networks},
  author={Lorenzo Vecci and Paolo Campolucci and Francesco Piazza and Aurelio Uncini},
  journal={Proceedings of International Conference on Neural Networks (ICNN'97)},
  year={1997},
  volume={1},
  pages={260-265 vol.1}
}
In this paper, we study the properties of neural networks based on adaptive spline activation functions. Using the results of regularization theory, we show how the proposed architecture is able to produce smooth approximations of unknown functions; to reduce hardware complexity a particular implementation of the kernels expected by the theory is suggested. This solution, although sub-optimal, greatly reduces the number of neurons and connections as it gives an increased expressive power to… CONTINUE READING

Figures, Tables, and Topics from this paper.

Citations

Publications citing this paper.
SHOWING 1-7 OF 7 CITATIONS

Speech and Speaker Recognition System Using Artificial Neural Networks and Hidden Markov Model

VIEW 1 EXCERPT
CITES METHODS

Recurrent Neural Networks for Prediction

VIEW 1 EXCERPT
CITES BACKGROUND

Fast complex adaptive spline neural networks for digital signal processing

References

Publications referenced by this paper.
SHOWING 1-9 OF 9 REFERENCES

Regression Modeling in1 Back-Porpagation Projection Pursuit Leantug

  • J. Wang, S. Lay, M. Maecler, R D.Martin, J. Schimert
  • IEEE Trans. on Neural Networks
  • 1994

Neural networks with digital LUT activation functions

Artificial Neural Networks with Adaptive Polynomial Activation Function

  • J. E. Moody, S. J. Hanson, +3 authors M. Zenobi
  • 1992