The term AI covers many sub-fields, including knowledge representation, neural networks, genetic algorithms, machine vision, and genetic programming. I've had the good fortune to work with both neural nets (NNs) and genetic programming (GP).
An example of the NN work I've done is available here. As my senior project at UW-Madison I ran several sets of data representing visual stimulation through a process of lateral inhibition and measured the effect of that processing on recognition in a neural network.
Another project I worked on involved genetic programming, a relatively new addition to the field of AI. GP works by generating populations of small programs, guaging their fitness, and then allowing those programs to reproduce and mutate. The end result is an individual or group of individual programs that are remarkably well suited to their environment, as defined by the fitness function.
The GP project I engaged in was an evaluation of the use of memory in the classic computer game Hunt the Wumpus. The paper I presented showing the results can be found here.
|Some AI web pages:||AAAI||MIT Media Lab||Genetic-Programming.org||AI at UW-Madison|