The Guardian has an article containing short interviews with Peter Molyneux and Steve Grand, two people who have managed to put out commercial games (in one case arguably commercially successful as well) containing "real" AI. Interesting read. They're both essentially pushing the idea that as games get even prettier, the stupidity of current game "AI" will shine through more and more, and so the need for "real" AI will increase, not increase.
I say maybe. While Molyneux's games are a great source of inspiration it's possible that it and its likes will always constitute a niche market, and your average FPS, RTS or movie tie-in adventure will never benefit from a neural network or evolutionary algorithm. But I do hope that I'm wrong here.
Whichever the case, we can still use commercial games for academic research just the same, in order to help us understand natural and computational intelligence. And I think we should. Much more than today.
I do remember reading an article about the AI in Colin McRae Rally. The arcade mode drivers were trained using neural nets, which were then saved at a variety of points to correspond with easy, average and hard opponents. I (well google) managed to find it here. There's not a lot of detail but it's quite interesting.
ReplyDelete(There's some good stuff in the search results actually).
I think that half of the problem with game AI is the fact that the computer can stick exactly to its plan, without losing concentration or getting angry. So game designers spend their time coding in scripted mistakes. Obviously these mistakes look great the first time, but quickly become predictable.
This happens a lot in GTR; if you get close to a car, it'll move to the inside and defend, but this is just a decision to switch to another path around the corner. They also out brake themselves from time to time. It'd be great if these events were driven by outside events, excess pressure etc.