Alert icon
We're changing our privacy policy. This stuff matters.  Learn more  Dismiss

Super Cool Mario AI trained by Genetic Algorithm

Loading...

Sign in or sign up now!
Alert icon
Upgrade to the latest Flash Player for improved playback performance. Upgrade now or more info.
7,469
Loading...
Alert icon
Sign in or sign up now!
Alert icon

Uploaded by on Dec 9, 2009

Neural Network Mario Player trained by a Genetic Algorithm. The NN has 14 inputs, 20 hidden neurons and 5 outputs.

  • likes, 1 dislikes

Link to this comment:

Share to:

Uploader Comments (tws220584)

  • You can find a lot more info in my report on: twsandberg.dk/portfolio.aspx under the section: Super Mario AI Agents

  • The point is to show that it is possible to use GA to train ANN to play procedural content generated levels in platform game like Super Mario (it was a school project), and there is also a competition for Super Mario.

  • The final version ran 800 generations.

see all

All Comments (11)

Sign In or Sign Up now to post a comment!
  • What kind of inputs did you use?

  • This sorta stuff it awesome, thanks for sharing it.

  • You are using it to play Super Mario?! WTF is the point of that if it can be (on supercomputers) better than humans (yes I understand this system)

  • how cud it be?

  • my mind FUCKING exploded

  • Neado!

    How many gens did it take?

Loading...

Alert icon
0 / 00Unsaved Playlist Return to active list
    1. Your queue is empty. Add videos to your queue using this button:
      or sign in to load a different list.
    Loading...Loading...Saving...
    • Clear all videos from this list
    • Learn more