Soft Computing Techniques are used throughout many of my other research intrests (eg Data Mining, Cellular Automata, Cognitive Science).
A recent interest and still in its infancy is the automated generation of visual art. See Internet links.
I am using evolutionary algorithms as a framework to generate computer graphic images. (see example images below).
A problematic area is that of fitness functions.
A number of computer artists by-pass this problem by relying on human aesthetic judgement (whether explicit or implicit).
I am looking to develop a computer based aesthetic model (probably connectionist) as fitness function for the existing evolutionary framework.
This may eventually be linked to to the work on emotion and emergence.

Image 1: Automated generated from initial random sequence of simple operators
Sequence: logt( minus( logxor( (logcom y) (logcom (logxor 33 x))) logand( y (abs 245))))
Further GreyScale evolved image population Here

Expressions use combinations of the following:
 [abs expt inverse logt sin cos tan rand logcom]
 [plus minus multiplication division modulus remainder logand logor logxor]