Brief Summary
Dr. Murphy, created this "smart" Mario using a reasonably greedy weighting algorithm to condition his Mario. The goal of his algorithm was to make the Mario win the game all the while maintaining the ascetics his Mario to move as if a human were playing and not a machine. The weighting Algorithm "rewards" (positively conditions) Mario whenever he makes his Score, Lives, Level, and Current X-Position on the screen increase, and negatively conditions him for doing the opposite (with some exceptions).
If you are interested in more information about this neural achievement please click on the "Watch Mario Learn!" link in which Dr. Murphy explains all of his ideas and the intricacies of the program or go to our references page and use any of the bottom three links. One of the links even holds the source code for the program!
If you are interested in more information about this neural achievement please click on the "Watch Mario Learn!" link in which Dr. Murphy explains all of his ideas and the intricacies of the program or go to our references page and use any of the bottom three links. One of the links even holds the source code for the program!