Thread: Artificial Intelligence in Games Using Color Mapping

Is there a way to implement color mapping in games, so that AI bots aren't ridiculously stupid? In tactical games like ArmA2, it would make camouflage much more important, and allow for easier stealth maneuvering.
Are there any games that do this already? I like the field of Artificial Intelligence, so this interested me. And would shape recognition be necessary for color mapping to work? For example, instead of shooting at things that contrast with the background, (sun and clouds on sky, etc.) you'd shoot at player-shaped color patches.

No game (with the possible exception of some research demos) implements vision as we know it for its AIs for one simple reason: it is slow. Rendering a scene takes up a pretty good fraction of the computing power spent on each frame. Doing that for every AI at the same time would leave you with 1 frame per second if you're lucky. If you consider turning down the quality of their vision or cheating and giving the bots something like thermal vision, you might as well go back to easier and most likely better-to-play methods.

Besides the problem of speed, artificial vision is not very advanced last I looked, so you'd probably end up with a bot that's more artificially stupid rather than artificially intelligent. Humans are amazingly good at pattern matching so we think it's easy, but when we go and try to program it, it really isn't. The bots would shoot at random rocks that cast the right shadow, but not the player standing in that shadow. (Look up the neural net that tried to learn the shape of tanks.)

This story might be apocryphal, but it doesn't really matter. It is a perfect illustration of the biggest problem behind neural networks. Any automatically trained net with more than a few dozen neurons is virtually impossible to analyze and understand. One can't tell if a net has memorized inputs, or is 'cheating' in some other way. A promising use for neural nets these days is to predict the stock market. Even though initial results are extremely good, investors are leery of trusting their money to a system that nobody understands.

That's a valid point, but the story in question isn't true. And I agree that it'd be slow; there's no simple way I can think of to simplify it.

In games what matters is what the player sees. Anything going on behind the scenes isn't even cake, it's just icing. If you can take shortcuts that the (average) player won't be able to see, you should take them. The trick with game AI isn't to make something that is human-like, only something that seems human-like and the trick to that is that it mostly comes down to a balancing act between too easy and too hard. (It's not very difficult to program an unbeatable AI for many games.)

Yea, some forms of intelligence are pretty artificial, all right.