Combining Self Organizing Maps and Multilayer Perceptrons to Learn Bot-Behavior for a Commercial Computer Game
Christian Thurau, Christian Bauckhage, Gerhard Sagerer · PUB – Publications at Bielefeld University (Bielefeld University) · 2003
Traditionally, the programming of bot behaviors for commercial computer games applies rule-based approaches. But even complex or fuzzyfied automatons cannot really challenge experienced players. This contribution examines whether bot programming can be treated as a pattern recognition problem and whether behaviors can be learned from recorded games. First, we sketch a technical computing interface to a commercial game that allows rapid prototyping of classifiers for bot programming. Then we discuss the use of self organizing maps to represent manifolds of high dimensional game data and how multilayer perceptrons can model local characteristics of such manifolds. Finally, some experiments in elementary behavior learning are presented.