Discussion in 'News Updates' started by The_J, Jul 14, 2011.
Who you would then destroy because it wasn't designed for that game.
That computer will never be productive again.
From Today's paper:
The paper edition also had a screen shot from a Civ 2 game.
I know it's been a week since this post, but I just wish to reply.
That is exactly how the AI learned. You give it a set of moves and let it figure out winning combinations of moves. (I am simplifying things quite a bit though) That is in fact pretty much how Chess AI works. It's just that chess is a bit of a different game than Civ.
Teaching AI is not about teaching it some 'winning' tactic, it is about giving it some set of actions and some criteria for these actions and letting it figure things out on its own. That is how modern rudimentary AIs work really. You can provide them guidance, but the whole point is that they figure out the 'tactics' on their own.
In this case, the manual provided additional information that helped the AI make a set of decisions that more reliably led to winning against the game's internal AI.
In fact, knowing the rules is precisely what leads to knowing how to play. The rules, and the actions they allow are exactly the building blocks of play. Take a simple rule from chess: knights move in an L-shaped pattern, skipping over everything but their landing square. This is a rule. It does not give you any 'play' as you put it, but that rule creates an action -- moving the knight in a specific pattern. Now, take that rule and combine it with all the many other rules. They each create specific allowed actions. These actions are then combined by the players into the 'play'. Without the rules, there is no play.
Separate names with a comma.