Using stochastic game theory and Q-Learning to train an AI in the game of Connect4
- Currently one can play against a human player or an artificial agent which was trained on an algorithm that randomly picks columns. The winning rate of this algorithm is 80% but it is clear that the game has not been learned that extensively. - Use stochastic game theory, instead of Markov Decision Process to train agent to be on par with human players.-
Notifications
You must be signed in to change notification settings - Fork 0
TaTo0101/connect4
Folders and files
| Name | Name | Last commit message | Last commit date | |
|---|---|---|---|---|
Repository files navigation
About
Using stochastic game theory and Q-Learning to train an AI in the game of Connect4
Resources
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published