Hi, I'm currently taking a course in AI and to complete the course we have to write a report. In the report were supposed to write examples of different AI techniques and then analyse them, write UML and component diagrams and so forth. The task isn't to think of the examples as much as to code them yourself and realize how they work.

The problem is that my preference is to make very small examples where I need to write very little code. But this far I haven't been able to come up with an example for the ANN part of the report.
This far I've been thinking of using a perceptron to determine which note a tone is. I've previously experimented with making my own sound waves with sdl_mixer so that part shouldn't be to difficult.
I would like to, if possible, avoid evolution and instead use back propogation, since I've already made an example with the evolution algorithm.

Is there a simple yet descriptive example out there that I haven't thought about?

EDIT:
Figured it out. I'm going to make a perceptron for a vendor that either accepts or declines an offer. Lika a haggling AI. Think it will be simple as its just numbercrunching.

I'll leave it to others to help you out on the ANN side, but I have a general question... is this a game AI course?

Well, I'm studying game programming. but this specific course is just AI in general.

Well, I'm studying game programming. but this specific course is just AI in general.

What school?

The simplest example of ANN is also the most important: A single neuron with linear activation function (a.k.a. linear regression). The next one in both orders is a single neuron with a sigmoid for activation function (a.k.a. logistic regression). In both of those cases, there are much better training algorithms than gradient descent or genetic algorithms.