Neural Networks with Higher Level Architecture for Bipolar Device Modeling

  • A. Plebe
  • A. M. Anile
  • S. Rinaudo
Conference paper
Part of the Mathematics in Industry book series (MATHINDUSTRY, volume 1)


Neural Networks (NN) are potential alternative to semiconductor modeling for circuit simulation, in situations where physical modeling becomes critical. To cope with the complex behavior of a state-of-the-art bipolar device, a particular NN architecture has been developed, referred here as Higher-Level, where the main neurons are instances of differential equations, and other neurons are responsible for the coefficients of such equations. Unfortunately this type of neural architecture is difficult to train and even the most sophisticate methods often fail to converge to an acceptable error. The strategy here presented essentially reduces the problem of training the higher-level NN to model the bipolar device in all its working conditions, to the training of simpler auxiliary networks, each working at a single DC bias point.


Bipolar Transistor Device Modeling Neural Network Architecture Neural Architecture Dynamic Neuron 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    G. Creech, J. Zurada, and P. Aronhime. Feedforward neural networks for estimating IC parametric yield and device characterisation. In Proc. of the IEEE Int. Symposium on Cyrcuits and Systems, Seattle, WA, 1995.Google Scholar
  2. 2.
    S.E. Fahlman Faster-learning variations on back-propagation: an empirical study. In D. Touretzky, Hinton G. E., and Sejnowski T. J., editors, Proc. of the 1988 Connectionist Models Summer School, San Mateo, CA, 1988. Morgan Kaufmann.Google Scholar
  3. 3.
    R.A. Jacobs. Increased rates of convergence through learning rate adaptation. Neural Networks, 1 (4): 295–307, 1988.CrossRefGoogle Scholar
  4. 4.
    B. L. Meijer, Peter. Neural Network Applications in Device and Subcircuit Modelling for Circuit Simulation. PhD thesis, Technical University of Eindhoven, 1996.Google Scholar
  5. 5.
    P. Meijer. Neural networks for device and circuit modelling. In Proc. of the Third Int. Workshop on Scientific Computing in Electrical Engineering, Warnemünde, Germany, 2000.Google Scholar
  6. 6.
    J. Nocedal and S. J. Wright. Numerical Optimization. Springer-Verlag, New York, 1999.zbMATHCrossRefGoogle Scholar
  7. 7.
    A. Plebe, A. M. Anile, and S. Rinaudo. Sub-micrometer bipolar transistor modeling using neural networks. In Proc. of the Third Int. Workshop on Scientific Computing in Electrical Engineering, Warnemünde, Germany, 2000.Google Scholar
  8. 8.
    W. L. Price. Global optimization algorithms for a CAD workstation. Journal of Optimization Theory and Applications, 55: 133–146, 1987.MathSciNetzbMATHCrossRefGoogle Scholar
  9. 9.
    K.-G. Rauh. A table model for circuit simulation. In Proc. of the Twelfth European Solid-State Circuit Conference, pages 211–213, Delft, The Netherlands, 1986.Google Scholar
  10. 10.
    A. J. Shepherd. Second-Order Methods for Neural Networks. Springer-Verlag, New York, 1997.CrossRefGoogle Scholar
  11. 11.
    T. Shima. Device and circuit simulator integration techniques. In W. L. Engl, editor, Process and device modeling, pages 433–459, Amsterdam, The Netherlands, 1986. North Holland.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2002

Authors and Affiliations

  • A. Plebe
    • 1
  • A. M. Anile
    • 1
  • S. Rinaudo
    • 2
  1. 1.Department of Mathematics and InformaticsUniversity of CataniaCataniaItaly
  2. 2.T MicroelectronicsCataniaItaly

Personalised recommendations