[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[microsound] Artificially Evolved Sounds





Dear List,


I'm a student and my major is artifical intelligence.
I'm looking for programmers who are intrested in joining me to try evolving sounds artificially according to the following plan. We
get sound fragments and classify them on a given scale. The value of a sound fragment come from our intuition how much we like the particular sound. After we have enough training samples we approximate/learn this function, for this purpose some kind of neural network or support
vector machine method could be used. We can test our function by giving it samples outside the training set and compare the value with our intuition. This function could be used as a fitness function in a genetic algorithm. The GA method would optimize (evolve) sounds that we like more than the sounds in the starting population. As far I as see there are a lot of opened question. To name a few we must find a capable representation of the sounds (spectral based for example.), find the learning method
of the fitness function and of course specify the details of the GA...


There will be a software art conference (the Readme 100) in Nov of this year:

"Readme festival in the year 2005 aims at supporting the production of software art projects and texts critically engaging with software art. Readme 100 will support up to 6 projects and up to 6 articles on the competition basis. Each project will get a budget from 500 to 3000 euros
....
Proposals for projects and texts should be sent to og {at} dxlab.org and inke.arns {at} hmkv.de no later than August 8."
(http://readme.runme.org/)


It would be nice to apply for support but I need programmers so that
we could able to implement the software together. If anyone of you are intrested in this please contact me offlist.
I like to ask you to decide rapidly because the proposal's
deadline is Aug.8 !!!



Regards, Gerofi B.

---------------------------------------------------------------------
To unsubscribe, e-mail: microsound-unsubscribe@xxxxxxxxxxxxx
For additional commands, e-mail: microsound-help@xxxxxxxxxxxxx
website: http://www.microsound.org