Neural networks are increasingly popular in geophysics. Because they are un
iversal approximators, these tools can approximate any continuous function
with an arbitrary precision. Hence, they may yield important contributions
to finding solutions to a variety of geo physical applications.
However, knowledge of many methods and techniques recently developed to inc
rease the performance and to facilitate the use of neural networks does not
seem to be widespread in the geophysical community. Therefore, the power o
f these tools has not yet been explored to their full extent. In this paper
, techniques are described for faster training, better overall performance,
i.e., generalization, and the automatic estimation of network size and arc
hitecture.