In the old days of gaming, I’m sure simple switch/case statements (in a sense) would have done just fine for most of the game “AI.” However, as games have become increasing complex, especially at the 3d leap, more complex algorithms are needed. My question is, are actual machine learning algorithms (like reinforcement learning) used in game AI at this point? Or is that still mostly only in research projects at universities (which I have been exposed to)?
If not actual machine learning algorithms, then what is driving bleeding edge commercial game AI? Is it simply highly complex but static (non-ML) algorithms that are able to cover most of the possibilities? And if so, what actual types of algorithms are used?
I’ve always been curious about this, thanks!
Edit: After thinking about it some more I can further clarify a bit. How do the agents in the game make decisions? If they are not using actual learning algorithms real-time, was a learning algorithm perhaps used in the development stages to produce a model (static algorithm), and that model is then used to make decisions in the game? Or was a static algorithm for decision making hand-coded in a sense?
Hi, In my opinion and I’m nowhere near being an expert in AI but based on my experience in game development beside playing a lot of games, I don’t think machine learning in games is commercially a viable option from a game design standpoint, despite the fact that it’s surly possible technological-wise.
Yes there are some rare instances where real-time machine training has taken place such as in Forza Motorsport (Drivatar) and pre-trained AI like in City Conquest, but in those instances the game as a whole doesn’t rely completely on the AI, it’s more like having a cherry on top!
Predictable gameplay behavior is necessary for story telling, game mechanics and progressive difficulty or you can’t simply guarantee a unified gameplay experience for all players otherwise due to different AI outcomes.
In classic (non-ML) AI, one of the ways agents make decisions is by planning ahead of time all possible solutions based on real-time environmental info and then take the best solution or shortest path for instance if we are talking about path finding algorithms for example, which is referred to as “Deterministic AI” the results are always certain and not just predictions like in “Non Deterministic AI”.