Have anyone got the subject? Or just theory of creating a NN operated programs? - need help...
Posted on 2001-09-19 07:12:02 by IIS
The best link I could point you to is CyberLife . However it seems to be down at the moment, so do a search for a article called "An exercise in creation". Its written by the creator of creatures and its a great intro into the theory.

There aren't going to be many examples on this subject, your best bet is to learn to write your own NN and then learn how to write a program for it to control. It can't be done the other way around.

The problem is that NN aren't anything in paticular. Secondly their all trained very differently, though the more used is back propagation it seems.

If in your searchs you comes across some intresting articles I encourage you to post links here. I myself have found introductions to various NNs hard to find.

Best of luck. :alright:
Posted on 2001-09-19 12:33:14 by Eóin
I know how NNs work but I've only got the Hebb learning algoritm. Anyone who knows how Backpropogation works? I've used genetic algorithms but they are quite hard to get working.
Posted on 2001-09-20 01:03:55 by gliptic
I've studied BPNN in my university course of "Artificial and Natural Intelligent Systems", but all my courses lessons are in Italian and hand written (sorry... if they were electronic I would have posted them).

I think a good starting point could be:

http://www.generation5.org/bp.shtml

and it's links.

Also... if you'd like to see a source code I found this one (I didn't look at it, sorry... I cannot say how it is done):

http://www.geocities.com/CapeCanaveral/1624/bpn.html

Hope to be usefull,

Daniele
Posted on 2001-09-22 05:45:18 by dguzz
I knew absolutely nothing about NN's at 1am this morning. Your topic intrigued me, however. One Google search, 13 hours, and 4 liters of pepsi later find me slightly more knowlegable.

http://www.google.com/search?=neural+network+implementation


There are even some very general articles out there for complete NN newbies such as myself. An article at IBM is noteworthy if even just for its Links section. That page has many example - coded in Python though. Also an interesting article on back-propogation there.

http://www-106.ibm.com/developerworks/linux/library/l-neural/

I don't know if this helps at all, being a NN newb it was hard for me to prioritize and parse

:tongue:

Good luck!
Posted on 2001-09-22 14:38:05 by Mark
You can find some great articles on Neural Network programming in reference to Artificial Intelligence, and particularly with respect to game enemies which can learn without memory per se, at the http://gamedev.net tutorial pages :) I particularly enjoyed the article on FuzzyLogic Neural Networks, quite clearly written.
Posted on 2001-09-29 15:22:37 by Evil Homer
Many thanks to all who answered to my question!

I found Generation5 to be a great source of NN-related information!
Posted on 2001-10-01 02:06:15 by IIS