We have created a UI on a web page that one can run locally to play against the agent. Two agents were trained using the AlphaZero algorithm, one for a 15x15 board and one for a 10x10. The generated models were exported and are queried to make a move after each move made by the user. Both are available for play on the webpage. The 10x10 is stronger due to a decreased time needed to train it.

Python3, Flask, Keras, numpy, Theano, Tensorflow

1. Set the flask entrypoint to mod_main.py by using export FLASK_APP=mod_main.py
2. Navigate to webapp/app
3. Start the server with flask run

• Shruti Alavala
• Jared Gruneiro
• Hannah Reed
• Mya Rios
• Parker Timmerman
• Michael Wynne