Well here I was thinking, you guys would love the topic, and not even one response in 10 hours ;p...
Could it be that you guys want me to explain NEAT a little bit more so you can then comment?
Well let us try...
NeuroEvolution of Augmenting Topologies (NEAT): is a genetic algorithm for the generation of evolving artificial neural networks (a neuroevolution technique) developed by Ken Stanley in 2002.
It is based on applying three key techniques: tracking genes with history markers to allow crossover among topologies, applying speciation (the evolution of species) to preserve innovations, and developing topologies incrementally from simple initial structures ("complexifying").
The NEAT approach begins with a perceptron-like feed-forward network of only input neurons and output neurons. As evolution progresses through discrete steps, the complexity of the network's topology may grow, either by inserting a new neuron into a connection path, or by creating a new connection between (formerly unconnected) neurons.
An extension of NEAT was devised in 2003 (rtNEAT) by Stanley that allows evolution to occur in real time, the basic idea is to put the population under constant evaluation with a "lifetime" timer on each individual in the population. When a network's timer expires its current fitness measure is examined to see whether it falls near the bottom of the population, and if so it is discarded and replaced by a new network bred from two high-fitness parents. A timer is set for the new network and it is placed in the population to participate in the ongoing evaluations.
The first application of rtNEAT is a video game called Neuro-Evolving Robotic Operatives, or NERO. In the first phase of the game, individual players deploy robots in a 'sandbox' and train them to some desired tactical doctrine. Once a collection of robots has been trained, a second phase of play allows players to pit their robots in a battle against robots trained by some other player, to see how well their training regimens prepared their robots for battle.
HyperNEAT: is specialized to evolve large scale structures. It was originally based on the CPPN theory and is an active field of research.
Content-Generating NEAT (cgNEAT): evolves custom video game content based on user preferences. The first video game to implement cgNEAT is Galactic Arms Race, a space-shooter game in which unique particle system weapons are evolved based player usage statistics.[3] Each particle system weapon in the game is controlled by an evolved CPPN, similarly to the evolution technique in the NEAT Particles interactive art program.
Implementations: Stanley's original (link:
http://nn.cs.utexas....ew.php?SoftID=4 ) and rtNEAT (link:
http://www.cs.utexas.../keyword?rtneat ) for C++
All this content was pasted from the wiki here's the link:
http://en.wikipedia....ting_topologies
From experience I can tell you this (I've been looking all over the place for it's implementations, program's that use it and so on...) I've used alot of programs that use rtNEAT (if you want I'll tell you the names, or even send to you the programs I have) and it's a beautifull thing to see believe me, the computer is a little bit slow to implement complex changes but once done, they are the best you can get.
I believe this could only be implemented in NWN not in real time on the servs (it would suck, the computer needs to start with basic things to do first, like approach enemy, or run from enemy) but by training the NPC's/monster's before battle and then putting them on the serv.
Another thing I can think could be done, since NEAT uses sensor's to see how to act, I believe we could use the sensors on a serv in real time, and then when training the NPC's put the things he learned form the sensors so he would learn from them.
Well here's my thought, hope you liked it, sorry if it's too long. ;p
PS: Sorry I don't know how to put link's just by clicking the words...
