Hey everyone,
Just to give you a quick update: My plan on creating a neural network
for SuperTux is still up to date.
I finished my seminar paper (If you're interested and proficient in
german: https://www.dropbox.com/s/y18t3b4nd3079sb/Seminararbeit.pdf?dl=0
, includes some small mistakes about neural network stuff) about neural
networks, currently I'm working on integrating a NEAT implementation to
the SuperTux source.
For now, the goal is to create a prototype that is capable of learning
basic game mechanics in an arbitrary amount of time.
In the master thesis, which is supposed to follow the current paper
("project work" or "Projektarbeit"), I intend to run SuperTux headless
so evolution can work at a way faster pace and take advantage from
faster and parallel hardware, among other things.
I'll keep you guys updated, maybe I'll shoot another mail if questions
occur.
Thanks!
Best Regards,
Christoph Kuhfuß
_______________________________________________
Supertux-Devel mailing list
Supertux-Devel@lists.lethargik.org
http://lists.lethargik.org/listinfo.cgi/supertux-devel-lethargik.org