by gsetRL
See inside
Notes and Credits

This is a reinforcement learning adaptation of the Harry Potter Snitch game. The goal is for the Harry Potter figure to grab the snitch. When the game begins, the agent explores, randomly using the grabbing command. It misses the snitch mostly, but it averages the first 5 hits into an ideal distance away from the snitch to grab. After 5 hits, it has learned the ideal distance and will not miss the snitch again.

Shared: 1 Aug 2009 Modified: 1 Aug 2009
Favorite this project 0
Love this project 0
Total views 243
View the remix tree  1
More projects by gsetRL