We implemented the well-known board game Connect Four in order to demonstrate the capabilities of machine learning in games.
AlphaZero is a general purpose deep reinforcement algorithm that uses Monte Carlo tree search to represent states and transitions of games. AlphaZero can be used for a rich variety of games. Adapting the algorithm to other board games is relatively straight forward. For example, a good article how this works can be found here.
The game room was made with our self developed 3D graphics framework which at its core uses three.js.
Click here in order tho starth this game in full-screen mode.