> Check out this timestamp to watch the machine "cheat": https://youtu.be/xOCurBYI_gY?t=9m55s
> Researcher's site about the project: http://www.cs.cmu.edu/~tom7/mario/
> The Paper: The First Level of Super Mario Bros. is Easy with Lexicographic Orderings and Time Travel...after that it gets a little tricky.: http://www.cs.cmu.edu/~tom7/mario/mario.pdf
The atari version even on the hard setting was almost fatally dumbed down to be mindless. The enemies were just way dumber than in the arcade version. The game really didn't even feel like Qbert.
With just a little practice, one could play on a single life for as long as desired. Similar to Asteroids on Atari.
Wow that's quite a jump to make
The title doesn't say that.
>It’s important to note, though, that the agent is not approaching this problem in the same way that a human would. It’s not actively looking for exploits in the game with some Matrix-like computer-vision.
[1] https://wiki.lesswrong.com/wiki/Wireheading
Another lesson: Evolutionary algorithms are really hard to control. Using neural networks developed through evolutionary algorithms means that we are employing a mostly opaque (though not entirely black) box created by a mechanism we can't mentally keep track of in detail. Hope that they are not deployed to control any critical systems until we get a much better grasp of them.
In this case one possible workaround to "cheating" would be to reduce the control precision, add some jittering to control inputs or change the goal function. But I'd say if it's being done solely with using the intended controls it's not cheating (as opposed to changing memory or using a debug 'cheat code').
Still, even in real sports some "cheating" is allowed (see Fosbury Flop)
From Wikipedia https://en.m.wikipedia.org/wiki/Gamesmanship
“Gamesmanship is the use of dubious (although not technically illegal) methods to win or gain a serious advantage in a game or sport. It has been described as "Pushing the rules to the limit without getting caught, using whatever dubious methods possible to achieve the desired end".”
To understand the concept of cheating, and to discuss what is cheating, requires an entirely higher cognitive capability.
Besides, how would you even tell the difference between a bug in the simulation and legitimate physics? I mean, look at electron tunneling.
My happiness won't change, but I would be excited.
If we are indeed in a simulator, then I would be compelled to create or join an effort to attract the attention of a being outside the simulator. Not for worship, but discourse.
To be able to communicate with something outside of what we had perceived as reality, and would be no less real, would be an amazing opportunity.
It literally-and-metaphorically can't see behind the UI to the internal state of the game.
I'd be disappointed if an AI chess computer invariably let me know that I had a better path to winning the game than it did.
Do you play on chess.com? That seems to attract unsporting players.
https://www.chess.com/forum/view/general/request-for-disabli...