Mario Kart MB - Print Page

Mario Kart MB
https://www.mariokart64.com/cgi-bin/yabb2/YaBB.pl
Mario Kart >> Super Mario Kart >> AI playing Super Mario Kart!
https://www.mariokart64.com/cgi-bin/yabb2/YaBB.pl?num=1515060383

Message started by KVD on 01/04/18 at 02:06:23

Title: AI playing Super Mario Kart!
Post by KVD on 01/04/18 at 02:06:23

See topic title and just watch the video!

[media]http://www.youtube.com/watch?v=HyJRfDaDLq4[/media]

I find this development very, very interesting (and as with everything AI based, also a little scary potentially) :P
It says that the AI is learning to play SMK based on the inputs that the gamer has provided (so it's mimicking + trial and error). I strongly wonder what would happen if we fed inputs of a top player into the initial states... [smiley=ninja.gif]

The AI and video are created by a famous name in gamer country, Seth Bling (https://en.wikipedia.org/wiki/SethBling). He also developed the credits warp in Super Mario World and actually build a working Atari computer within Minecraft! :o

Title: Re: AI playing Super Mario Kart!
Post by Fabrice Baro on 01/04/18 at 05:55:20

Very interesting! I had seen MarI/O in TT before, this seems like a better "driver".
Imho the initial input isn't too important. If the program is configured correctly it will get better over time (i.e. trying over and over again and getting better and better results).

Title: Re: AI playing Super Mario Kart!
Post by KVD on 01/04/18 at 06:05:04

That's very true, depending on how well it is configured and how many iterations are done. Potentially this method will deliver the absolute perfect TASes and new techniques that humans may not even be able to execute. But a lot will depend on how the configuration is done though. It's probably very difficult to get it right. Arriving at complex techniques by randomly (unguided) trying out button sequences is something that WILL take a very long time (same problem with the evolution of complex life forms!). The result it gets should then be presented back to the program as a sort of natural selection principle. Therefore, the initial inputs may not be that trivial after all if you want to greatly reduce the number of iterations needed (intelligent design :P).

I find the parallel with the evolution of life very interesting. Evolution by natural selection (which is effectively the neural networks approach) takes an extravagant amount of iterations, whereas guided evolution (humans editing genomes with purpose and direction) is potentially very fast, but may not produce the most optimal results. The combo of the two is most likely the most powerful / efficient.

Title: Re: AI playing Super Mario Kart!
Post by Fabrice Baro on 01/05/18 at 12:25:13

456F7C6B620E0 wrote:

I find the parallel with the evolution of life very interesting. Evolution by natural selection (which is effectively the neural networks approach) takes an extravagant amount of iterations, whereas guided evolution (humans editing genomes with purpose and direction) is potentially very fast, but may not produce the most optimal results. The combo of the two is most likely the most powerful / efficient.

^ Yes! Evolutionary timescales are extremely long, hard for most (including myself) to truly comprehend. This allows for a *lot* of tries.

Also many life forms have embedded mechanisms that allow them to change quickly either during a lifetime or in just a few generations (these capabilities are the result of "plain" natural selection of course):

https://en.wikipedia.org/wiki/Phenotypic_plasticity
https://en.wikipedia.org/wiki/Polyphenism

Title: Re: AI playing Super Mario Kart!
Post by Etch on 01/05/18 at 18:46:06

Creating AI to surpass TAS should be the ultimate goal!

God help us!!!

http://i.imgur.com/xLhDmn3.png

Title: Re: AI playing Super Mario Kart!
Post by Chrono Krysster on 01/28/18 at 08:04:30

Indeed that is very interesting, please Karel let me add MariFlow to the TensorKart & MarI/O thread (http://www.mariokart64.com/cgi-bin/yabb2/YaBB.pl?num=1483724001/0#0) for more coverage on both topics.