A browser-based, educational game that demonstrates the way that artificial intelligence makes decisions through a neural network.
The goal of the game is to pick up the final stick from the pile. Each turn, the player will pick up between 1 and 3 sticks. At the conclusion of the game, the AI (Nathaniel) will update his neural network based on the moves he made and the result of the game. His neural net can be seen by the player on the right side of the screen.
With the ability to simulate 0, 50, 100, and 500 games before the start of a game, users can see Nathaniel learn through play.
To run the game locally, you will need to run a local webserver. WAMP, MAMP, XAMMP all work great for whatever platform you may be in.
Clone the repository into your webserver's default folder.
If you wish to host the game, you will follow similar steps as above for whatever hosting service you are using.
If hosting locally, navigate to localhost:<port> in the browser of your choice.
Else, navigate to the url where you have hosted the project.
- Some assets are lacking or missing (specifically sliders)
- Simulations run much slower on the Galaxy tablets
- AI training model optimization
- Analytics and visualization of AI learning
- Sandbox mode (ability to edit values in the neural net)
-
Lia Dawson
- GitHub: ldawson02
-
William Taylor
- GitHub: wtaylor45
- Email: [email protected]
-
Greg Mohler
- GitHub: gmohler213
- Email: [email protected]
-
Rachel Muzzelo
- Email: [email protected]
MIT - see license file