Connect-Four Learning AI
Made: December 2013 • Dev Progress: 100%
Click the various options to run the learning algorithm.
About: This program uses the Monte Carlo method of reinforcement learning to build a set of policies off which it plays connect four. The algorithm works with a variety of board dimensions and connect-number goals.
You can even play against the AI once it has learned enough!
The source code can be viewed and downloaded here.
Read my technical write up for making this program (PDF Download):
This file requires Adobe Flash Player (Standalone) to run, or a browser with Flash Player installed.