• DocumentCode
    1722725
  • Title

    Computation of a nonlinear squashing function in digital neural networks

  • Author

    Havel, Vladimír ; Vlcek, Karel

  • Author_Institution
    Dept. of Comput. Sci., Tech. Univ. of Ostrava, Ostrava
  • fYear
    2008
  • Firstpage
    1
  • Lastpage
    4
  • Abstract
    The paper deals with implementing nonlinear sigmoid-like squashing functions in FPGA based neural networks. Five piecewise linear (PWL) approximation techniques and a look-up table are compared. It is examined their performance in terms of recall delay, area requirements, and accuracy in two FPGA neural networks.
  • Keywords
    approximation theory; field programmable gate arrays; neural chips; piecewise linear techniques; table lookup; FPGA; digital neural network; look-up table; nonlinear squashing function; piecewise linear approximation; recall delay; Artificial neural networks; Computer networks; Field programmable gate arrays; Logic arrays; Neural network hardware; Neural networks; Parallel processing; Piecewise linear approximation; Piecewise linear techniques; Table lookup;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Design and Diagnostics of Electronic Circuits and Systems, 2008. DDECS 2008. 11th IEEE Workshop on
  • Conference_Location
    Bratislava
  • Print_ISBN
    978-1-4244-2276-0
  • Electronic_ISBN
    978-1-4244-2277-7
  • Type

    conf

  • DOI
    10.1109/DDECS.2008.4538748
  • Filename
    4538748