A Fast, Compact Approximation of the Exponential Function

N. N. Schraudolph. A Fast, Compact Approximation of the Exponential Function. Neural Computation, 11(4):853–862, 1999.

Download

175.3kB	71.2kB	157.8kB

Abstract

Neural network simulations often spend a large proportion of their time computing exponential functions. Since the exponentiation routines of typical math libraries are rather slow, their replacement with a fast approximation can greatly reduce the overall computation time. This paper describes how exponentiation can be approximated by manipulating the components of a standard (IEEE-754) floating-point representation. This models the exponential function as well as a lookup table with linear interpolation, but is significantly faster and more compact.

BibTeX Entry

@article{Schraudolph99,
     author = {Nicol N. Schraudolph},
      title = {\href{http://nic.schraudolph.org/pubs/Schraudolph99.pdf}{
               A Fast, Compact Approximation of the Exponential Function}},
      pages = {853--862},
    journal = {\href{http://neco.mitpress.org/}{Neural Computation}},
     volume =  11,
     number =  4,
       year =  1999,
   b2h_type = {Journal Papers},
  b2h_topic = {Other},
   abstract = {
    Neural network simulations often spend a large proportion of their time
    computing exponential functions.  Since the exponentiation routines of
    typical math libraries are rather slow, their replacement with a fast
    approximation can greatly reduce the overall computation time.  This
    paper describes how exponentiation can be approximated by manipulating
    the components of a standard (IEEE-754) floating-point representation.
    This models the exponential function as well as a lookup table with
    linear interpolation, but is significantly faster and more compact.
}}