Abstract
Neural networks, applications of the computational model inspired
by the human brain, have proven an invaluable tool in solving a
wide variety of input-output mapping problems.
Scientific visualization software is used to understand both the
learning and computational processes involved.
This project presents an interactive visualization tool designed
to overcome several limitations of existing packages.
Using a genetic algorithm to evolve problem solutions,
NVIS provides graphical representations of both the
individual networks and the genealogical relationships within the
population.